View in EDS HTML Full Text PDF Full Text

Mapping Academic Perspectives on AI in Education: Trends, Challenges, and Sentiments in Educational Research (2018-2024)

Saved in:
Bibliographic Details
Title: Mapping Academic Perspectives on AI in Education: Trends, Challenges, and Sentiments in Educational Research (2018-2024)
Language: English
Authors: Ji Hyun Yu (ORCID 0000-0001-9648-2582), Devraj Chauhan, Rubaiyat Asif Iqbal, Eugene Yeoh
Source: Educational Technology Research and Development. 2025 73(1):199-227.
Availability: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
Peer Reviewed: Y
Page Count: 29
Publication Date: 2025
Document Type: Journal Articles
Information Analyses
Descriptors: Artificial Intelligence, Technology Uses in Education, Educational Trends, Trend Analysis, Educational Research, Technology Integration, Ethics, Educational Policy, Policy Formation, Evidence Based Practice, Decision Making
DOI: 10.1007/s11423-024-10425-2
ISSN: 1042-1629
1556-6501
Abstract: How is the academic community conceptualizing and approaching the integration of AI in education, considering its potential, complexities, and challenges? This study addresses this fundamental question by employing a multifaceted approach that combines co-occurrence network analysis, latent Dirichlet allocation (LDA), and sentiment analysis on a corpus of abstracts from academic publications from 2018 to 2024. The findings reveal key themes in the scholarly discourse, including the centrality of ethical considerations, the impact of global events on AI adoption, and the practical applications of AI in educational management and policymaking. Moreover, the study identifies the main factors discussed in literature as influencing successful AI integration, the challenges and opportunities associated with AI in education, and the evolving academic perspectives on AI's role in educational settings. This comprehensive analysis of academic literature provides valuable insights into the current state of AI in education research, highlighting trends, challenges, and sentiments as they have evolved over time. By mapping the landscape of scholarly thought on this topic, this study aims to inform future research agendas, contribute to policy discussions, and provide a foundation for evidence-based decision-making in the development and implementation of AI technologies in educational contexts.
Abstractor: As Provided
Entry Date: 2025
Accession Number: EJ1462602
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwHec8xL6CVZsvcFF5jqu8yoAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDDFD52ZsN5L8l30syAIBEICBm80wxlb6BvHPnDxMz4R5rXKLdAJJQsy0AQp0h-qI_2g_nA-16v0YMAvyJxCiTxEYUmdsvPr_0SzIWWB_lWQJVqPzawmg1IAhOyHybMur7i88euN2OiSMJAfBou812R1wMgm0D94rWFbfwVALuGmnVTpqCqGtbRXZ7ay-Npb8hTWcs01a4H10jnmJwjjAIA-yZqiiwAIyKhCUWguo
Text:
  Availability: 1
  Value: <anid>AN0183751070;etr01feb.25;2025Mar19.03:45;v2.2.500</anid> <title id="AN0183751070-1">Mapping academic perspectives on AI in education: trends, challenges, and sentiments in educational research (2018–2024) </title> <p>How is the academic community conceptualizing and approaching the integration of AI in education, considering its potential, complexities, and challenges? This study addresses this fundamental question by employing a multifaceted approach that combines co-occurrence network analysis, latent Dirichlet allocation (LDA), and sentiment analysis on a corpus of abstracts from academic publications from 2018 to 2024. The findings reveal key themes in the scholarly discourse, including the centrality of ethical considerations, the impact of global events on AI adoption, and the practical applications of AI in educational management and policymaking. Moreover, the study identifies the main factors discussed in literature as influencing successful AI integration, the challenges and opportunities associated with AI in education, and the evolving academic perspectives on AI's role in educational settings. This comprehensive analysis of academic literature provides valuable insights into the current state of AI in education research, highlighting trends, challenges, and sentiments as they have evolved over time. By mapping the landscape of scholarly thought on this topic, this study aims to inform future research agendas, contribute to policy discussions, and provide a foundation for evidence-based decision-making in the development and implementation of AI technologies in educational contexts.</p> <p>Keywords: AI in education; Co-occurrence network; Topic modeling; Sentiment analysis</p> <p>Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.</p> <hd id="AN0183751070-2">Introduction</hd> <p>The rapid advancement of Artificial Intelligence (AI) in education presents both opportunities and challenges for the academic community as they explore the integration of these technologies into their teaching practices. AI provides a wide range of potential benefits, such as personalized learning paths (Huang et al., [<reflink idref="bib38" id="ref1">38</reflink>]), adaptive assessment methods (Bower et al., [<reflink idref="bib14" id="ref2">14</reflink>]), and human-like tutoring systems (Alkhatlan & Kalita, [<reflink idref="bib8" id="ref3">8</reflink>]). However, as we've experienced with technology integration in education over the past decades, the successful integration of AI in education is not simply about the technology itself (Ertmer et al., [<reflink idref="bib29" id="ref4">29</reflink>]); it requires a deep understanding of its impact on pedagogical approaches, ethical considerations, and learning outcomes.</p> <p>In recent years, there has been a surge in academic literature examining various aspects of AI in education (e.g., Berendt et al., [<reflink idref="bib11" id="ref5">11</reflink>]; Gillani et al., [<reflink idref="bib32" id="ref6">32</reflink>]; Holmes & Tuomi, [<reflink idref="bib35" id="ref7">35</reflink>]; Memarian & Doleck, [<reflink idref="bib53" id="ref8">53</reflink>]; Tahiru, [<reflink idref="bib69" id="ref9">69</reflink>]; Zhai et al., [<reflink idref="bib82" id="ref10">82</reflink>]). For instance, Berendt et al. ([<reflink idref="bib11" id="ref11">11</reflink>]) examined the implications of AI for fundamental human rights in educational contexts. Gillani et al. ([<reflink idref="bib32" id="ref12">32</reflink>]) worked on demystifying AI for educators, aiming to make complex concepts more accesible. Holmes and Tuomi ([<reflink idref="bib35" id="ref13">35</reflink>]) developed a typology of AI systems in education, offering a structured understanding of different AI applications. Memarian and Dolect ([<reflink idref="bib53" id="ref14">53</reflink>]) explored the role of embodied AI in learning, highlighting the interplay between technology, cognition, and physical experience. Tahiru ([<reflink idref="bib69" id="ref15">69</reflink>]) conducted a systematic literature review to analyze the opportunities, benefits, and challenges of AI in education. Zhai et al. ([<reflink idref="bib82" id="ref16">82</reflink>]) provided a comprehensive review of AI in education research from 2010 to 2020, categorizing studies into development, application, and integration layers. These diverse studies collectively demonstrate the multifaceted nature of AI in education research, spanning technical, ethical, pedagogical, and practical considerations.</p> <p>While these studies have significantly contributed to our understanding of AI in education, they primarily focus on specific aspects or provide broad overviews of the field. What is lacking is a comprehensive, data-driven analysis of how the academic discourse on AI in education has evolved over time, particularly in recent years when AI technologies have rapidly advanced. There is a need to understand not just what topics are being researched, but how key concepts in the field are interconnected, how thematic structures have developed, and how academic perspectives have shifted over time.</p> <p>To address this gap, our study aims to provide a comprehensive analysis of the academic discourse surrounding AI integration in educational settings, including emerging themes, key concepts, and evolving perspectives as reflected in published literature from 2018 to 2024. We focus on the intersection of AI applications in education and the role of educators in this technological integration. This focus is reflected in our data collection strategy, which employs key search terms related to human-AI collaboration in educational contexts. By targeting terms such as "Human-AI," "Educator-AI," "AI-enhanced teaching," and "pedagogical AI," alongside terms related to teaching strategies and pedagogical decisions. This targeted approach allows us to examine not only the technical aspects of AI in education but also the human element—how educators are adapting to, utilizing, and shaping AI technologies in their teaching practices. By doing so, we aim to provide a broad overview understanding of the practical implementation of AI in education and the changing role of educators in an increasingly AI-enhanced educational environment.</p> <p>By employing a multi-faceted approach that combines co-occurrence network analysis, latent Dirichlet allocation (LDA), and sentiment analysis on a corpus of academic publications, this research seeks to reveal the key concepts, thematic structures, and emotional perspectives that shape the discourse on AI in education. This text mining approach allows for a broad, longitudinal view of trends and themes in the academic discourse, offering unique insights into how the field of education is grappling with the rapid advancement of AI technologies.</p> <hd id="AN0183751070-3">Literature review</hd> <p></p> <hd id="AN0183751070-4">Artificial intelligence in education</hd> <p>The integration of artificial intelligence in educational settings is a significant shift in the approach to teaching and learning. Machine learning (ML) algorithms, as a key driver of AI, have revolutionized educational assessment and feedback mechanisms. Adaptive learning platforms, such as Khan Academy, utilize ML algorithms to analyze students' performance data and dynamically adjust the difficulty and content of learning materials to meet their individual needs (Perrotta & Selwyn, [<reflink idref="bib59" id="ref17">59</reflink>]). Teachers can use these platforms to monitor students' progress, identify areas of strength and weakness, and deliver targeted support as needed (Whitehill & Seltzer, [<reflink idref="bib75" id="ref18">75</reflink>]; Williams et al., [<reflink idref="bib76" id="ref19">76</reflink>]). By providing personalized recommendations and targeted interventions, these platforms empower students to progress at their own pace, reinforce concepts they are struggling with and facilitate their learning journey. In addition, ML-based plagiarism detection tools help educators perform academic integrity by identifying and addressing instances of plagiarism in students' work, fostering a culture of honesty and accountability in education (Hellas et al., [<reflink idref="bib34" id="ref20">34</reflink>]; Khalil & Er, [<reflink idref="bib44" id="ref21">44</reflink>]).</p> <p>Deep learning (DL), a subset of ML, has led to breakthroughs in educational research and content creation. One example is the development of intelligent tutoring systems (ITS), which use DL algorithms to model students' cognitive processes and adapt instructional strategies accordingly. ITS platforms like Carnegie Learning's MATHia analyze students' problem-solving approaches in real-time, providing personalized hints, scaffolding, and feedback to support their learning goals (King et al., [<reflink idref="bib47" id="ref22">47</reflink>]). Similarly, DL-powered content generation tools like OpenAI's GPT-3.5 enable educators to create interactive learning materials, generate quiz questions, and facilitate engaging classroom discussions, enhancing the richness and diversity of educational resources available to learners.</p> <p>Natural language processing (NLP), another subdiscipline of ML, has advanced significantly, enabling AI-powered virtual assistants to interact with students, answer their questions, and provide personalized support in real-time (Shaik et al., [<reflink idref="bib67" id="ref23">67</reflink>]). For example, language learning platforms like Duolingo utilize AI-driven chatbots to simulate conversation, correct pronunciation, and customize lessons to individual proficiency levels, providing students with a dynamic and immersive learning experience (Purgina et al., [<reflink idref="bib62" id="ref24">62</reflink>]). Additionally, advanced NLP models like OpenAI's GPT-3.5, which is based on deep learning architectures, have further transformed the way educational content is generated, allowing for the creation of more interactive and engaging learning materials.</p> <p>However, the integration of AI in education presents challenges related to data privacy, security, and ethics (Akgun & Greenhow, [<reflink idref="bib3" id="ref25">3</reflink>]). AI-powered systems collect and analyze vast amounts of student data, raising concerns about responsible use. To address these concerns, robust protections and transparent governance frameworks are needed to protect sensitive information and mitigate the risk of algorithmic bias or discrimination (Kordzadeh & Ghasemaghaei, [<reflink idref="bib48" id="ref26">48</reflink>]). Additionally, the digital divide remains a significant barrier to equitable access to AI-powered educational resources, particularly for underserved communities with limited internet connectivity or access to technology (Božić, [<reflink idref="bib15" id="ref27">15</reflink>]). Bridging this gap requires concerted efforts to address infrastructure gaps, promote digital literacy, and expand affordable access to hardware and internet services (Ehimuan et al., [<reflink idref="bib28" id="ref28">28</reflink>]). Despite these challenges, the transformative potential of AI in education is undeniable, offering new opportunities to enhance teaching effectiveness, personalize learning experiences, and empower students to succeed in the digital age.</p> <p>Identifying and addressing these research gaps is crucial to understanding the full potential and implications of AI in education, ensuring that its integration is guided by evidence-based practices and informed by a deep understanding of its impact on teaching and learning outcomes (Wang & Cheng, [<reflink idref="bib73" id="ref29">73</reflink>]). By examining the existing literature and synthesizing key findings, this study aims to contribute to the ongoing discourse on AI in education, addressing emerging trends, identifying areas for further investigation, and providing practical insights for educators, policymakers, and researchers.</p> <hd id="AN0183751070-5">Academic perspectives on AI in education</hd> <p>The integration of AI in education has sparked significant academic interest, as evidenced by recent systematic reviews (Celik et al., [<reflink idref="bib18" id="ref30">18</reflink>]; Cheng et al., [<reflink idref="bib20" id="ref31">20</reflink>]; Crompton & Burke, [<reflink idref="bib24" id="ref32">24</reflink>]; Ouyang et al., [<reflink idref="bib57" id="ref33">57</reflink>]; Zawacki-Richter et al., [<reflink idref="bib81" id="ref34">81</reflink>]). These reviews highlight AI's potential as a major driver for improving pedagogical practices while also acknowledging the multifaceted nature of its impact on education.</p> <p>Scholars emphasize AI's potential to enhance both personalized learning and operational efficiency in educational environments (Chiu et al., [<reflink idref="bib21" id="ref35">21</reflink>]). The academic discourse on AI in education encompasses various applications, including student evaluation and performance prediction, administrative tasks and classroom management, AI-assisted learning systems, and data-informed personalized learning. AI tools offer the promise of tailored learning experiences, potentially contributing to improved student outcomes (Ouyang et al., [<reflink idref="bib56" id="ref36">56</reflink>]).</p> <p>Holmes and Tuomi ([<reflink idref="bib35" id="ref37">35</reflink>]) developed a typology of AI in education (AIED) systems, describing different ways of using AI in education and learning. They argue that these systems are grounded in various interpretations of what AI and education could be, highlighting the diverse potential applications of AI in educational settings. Building on this, Gillani et al. ([<reflink idref="bib32" id="ref38">32</reflink>]) "unpacked the black box" of AI in education, offering a basic introduction to different methods and philosophies underpinning AI. Their work explores recent advances, applications to education, and key limitations and risks, providing a comprehensive overview of AI's potential in educational contexts.</p> <p>Despite the potential benefits, the literature also identifies several challenges in integrating AI into education. Researchers note that AI systems can sometimes make errors, and their effectiveness may depend on users' familiarity with the technology (Farrokhnia et al., [<reflink idref="bib30" id="ref39">30</reflink>]). A significant theme in the literature is the perceived lack of AI literacy among educators and the need for professional development in this area (Casal-Otero et al., [<reflink idref="bib16" id="ref40">16</reflink>]).</p> <p>Berendt et al. ([<reflink idref="bib11" id="ref41">11</reflink>]) highlighted the need to balance the benefits and risks of AI in education, particularly concerning fundamental human rights and freedoms of both teachers and learners. They called for embedding considerations of these benefits and risks into the development, marketing, and deployment of AI tools in education. This emphasis on ethical considerations is echoed throughout the literature, with scholars highlighting the importance of understanding not just the technical aspects of AI but also its social and ethical implications (Long & Magerko, [<reflink idref="bib50" id="ref42">50</reflink>]).</p> <p>The scarcity of learning opportunities related to AI in education is another challenge that may impede effective implementation (Ghamrawi et al., [<reflink idref="bib31" id="ref43">31</reflink>]; Sun et al., [<reflink idref="bib68" id="ref44">68</reflink>]). Velander et al. ([<reflink idref="bib72" id="ref45">72</reflink>]) identified key issues in the academic discourse, including ambiguous understanding of AI in educational contexts, limited knowledge about AI among education professionals, and a range of emotional responses to AI from excitement to skepticism.</p> <p>The literature reveals distinct challenges in creating AI literacy curriculum, particularly due to the lack of agreement on the definition of AI literacy (Casal-Otero et al., [<reflink idref="bib16" id="ref46">16</reflink>]). Some initiatives focus on equipping students with technical skills like programming and data analysis (Touretzky et al., [<reflink idref="bib70" id="ref47">70</reflink>]), while others highlight the importance of understanding the ethical and social implications of AI.</p> <p>Memarian and Doleck ([<reflink idref="bib53" id="ref48">53</reflink>]) explored the concept of embodied AI in education, examining the interplay between mind, body, and environment in AI-enhanced learning contexts. Their work highlights the need for a more holistic approach to AI in education, considering not just the technological aspects but also the physical and cognitive dimensions of learning.</p> <p>Zhai et al. ([<reflink idref="bib82" id="ref49">82</reflink>]) provided a comprehensive review of AI in education from 2010 to 2020, categorizing research questions into development, application, and integration layers. They identified emerging research trends such as the Internet of Things, swarm intelligence, deep learning, and neuroscience, while also highlighting challenges related to the inappropriate use of AI techniques and changing roles of teachers and students.</p> <p>In conclusion, while the potential of AI in education is significant, the academic discourse emphasizes the need for careful consideration of its implementation. This includes addressing issues of AI literacy among educators, ensuring ethical use of AI technologies, and considering the broader implications of AI integration in educational settings. Future research directions, as suggested by the literature, include exploring more holistic approaches to AI in education, addressing ethical concerns, and developing comprehensive AI literacy frameworks for both educators and students.</p> <hd id="AN0183751070-6">Purpose of the study</hd> <p>While existing literature provides valuable insights into AI in education, there's a lack of comprehensive, data-driven analysis examining how academic discourse has evolved, especially as AI technologies have rapidly advanced. Our study addresses this gap by employing network analysis, topic modeling, and sentiment analysis on a large corpus of academic literature, aiming to reveal patterns and trends in how the academic community conceptualizes and approaches AI integration in educational settings.</p> <p>Our investigation is driven by three critical research questions:</p> <p></p> <ulist> <item> What key concepts emerge in academic discussions about AI integration in educational settings, and how are these concepts interconnected?</item> <p></p> <item> How do thematic structures identified through latent Dirichlet allocation (LDA) describe the various aspects of AI integration in education as discussed in academic literature?</item> <p></p> <item> How has the academic perspective towards AI in education evolved over time, as reflected in sentiment analysis of published research?</item> </ulist> <hd id="AN0183751070-7">Methods</hd> <p>To produce a thorough and accurate review of existing research, we followed a detailed process informed by previous work on systematic literature reviews (Ifenthaler & Yau, [<reflink idref="bib41" id="ref50">41</reflink>]) and text mining (Bhutoria, [<reflink idref="bib12" id="ref51">12</reflink>]). The systematic review method allowed us to gather and evaluate all relevant studies, ensuring our review is exhaustive and unbiased in mapping academic perspectives on AI in education, including trends, challenges, and sentiments in the field. Text mining, on the other hand, enabled efficient processing and analysis of large volumes of textual data, facilitating the discovery of patterns, trends, and relationships within literature that may not be immediately apparent. These methods greatly improved our ability to achieve our research goals by providing a structured and data-driven basis for identifying the current state of AI applications in education and revealing areas for further investigation.</p> <hd id="AN0183751070-8">Systematic review process</hd> <p>Once we identified the goals and questions for our research, we developed a search protocol and prepared our team for a comprehensive search process. To align closely with our goal of understanding both the practical use and sentiment towards AI in education, we chose to focus our initial search on two areas: the application of AI in education contexts and the role of educators in this technological integration. The selected key search terms were carefully chosen to capture a broad spectrum of AI applications in education and educators' involvement. These terms include:("Human-AI" OR "Educator-AI" OR "AI-enhanced" OR "AI-powered" OR "teacher-AI" OR "pedagogical AI" OR "classroom AI" OR "AI in education" OR "AI in teaching" OR "AI in the classroom" OR "AI tutors" OR "teaching with AI")AND("teach*" OR "teaching strategies" OR "pedagog*" OR "classroom" OR "pedagogical decision*")</p> <p>We carefully defined our inclusion and exclusion criteria to ensure our review focused on the most relevant and current research. Specifically, we chose to include: (<reflink idref="bib1" id="ref52">1</reflink>) peer-reviewed journal articles published between 2018 and 2024 to capture the latest development and trends in AI within education, reflecting the rapid advancement in this field; (<reflink idref="bib2" id="ref53">2</reflink>) articles written in English, due to our team's proficiency and to ensure a broad reach in the global research community; (<reflink idref="bib3" id="ref54">3</reflink>) studies with available abstracts, crucial for our quick relevance check and as a corpus for co-occurrence network analysis, sentiment analysis, and topic; (<reflink idref="bib4" id="ref55">4</reflink>) research specifically examining how educator utilize AI systems or tools in educational settings; (<reflink idref="bib5" id="ref56">5</reflink>) articles discussing formal educational settings such as K-12, higher education, professional training, where the application of AI has a direct impact on teaching and learning processes; and (<reflink idref="bib6" id="ref57">6</reflink>) studies centered on practical applications of AI, prioritizing real-world usage over theoretical discussions, modeling, or review studies to ground our findings in tangible evidence.</p> <p>During our initial search, we found that certain terms attracted a high volume of theoretical and review articles not aligned with our emphasis on practical applications. Also, we specified our educational settings to focus on formal education settings. To refine our search and better focus on empirical research, we updated our key search terms:("Human-AI" OR "Educator-AI" OR "AI-enhanced" OR "AI-powered" OR "teacher-AI" OR "pedagogical AI" OR "classroom AI" OR "AI in education" OR "AI in teaching" OR "AI in the classroom" OR "AI tutors" OR "teaching with AI")AND("teach*" OR "teaching strategies" OR "pedagog*" OR "classroom" OR "pedagogical decision*")NOT("systematic review" OR "systematic literature review")AND ("K-12" OR "primary school*" OR "elementary school*" OR "middle school*" OR "high school*" OR "college" OR "professional" OR "training" OR "career development")</p> <p>This adjustment was based on insights from our practice search, which highlighted the need to more precisely target studies that describe how AI is being implemented by educators across different stages of education and professional development.</p> <p>From our searches in the selected databases, such as Google Scholar, Scopus, Web of Science, ERIC, and ACM Library, as well as a review of the top 20 journals in the field of educational technology as identified by Google,[<reflink idref="bib1" id="ref58">1</reflink>] we initially identified 425 peer-reviewed journal articles. To refine this list to those most relevant to our study, we first removed duplicates, ensuring each article was unique. Next, we evaluated each article for relevance to our specific search focus, by carefully reviewing their abstracts to assess its alignment with our research questions. Articles that did not focus on the direct application of AI in educational practices or were outside the scope of formal educational settings were excluded. Additionally, we prioritized empirical research over theoretical studies, reviews, or modeling papers, aiming to focus on articles that provide insights into actual use cases and experiences of educators with AI. Through this detailed screening and selection process, we filtered our initial dataset to a final collection of 269 publications that directly contribute to understanding the trends, challenges, and opportunities of AI use by educators in educational environments.</p> <p>Table 1 provides a summary of the key features of our final dataset, including the distribution of papers across different educational settings and key focus areas.</p> <p>Table 1 Summary of key features (n = 269)</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Educational setting</p></th><th align="left"><p>Number of papers</p></th><th align="left"><p>Key focus area</p></th></tr></thead><tbody><tr><td align="left"><p>K-12</p></td><td align="left"><p>37</p></td><td align="left"><p>Application of AI in primary, elementary, middle, and high school settings, enhancing teaching and learning processes</p></td></tr><tr><td align="left"><p>Higher education</p></td><td align="left"><p>68</p></td><td align="left"><p>Use of AI in colleges and universities, impact on teaching strategies, student engagement, and administrative processes</p></td></tr><tr><td align="left"><p>Professional training</p></td><td align="left"><p>15</p></td><td align="left"><p>Role of AI in professional training, career development, and professional development programs</p></td></tr><tr><td align="left"><p>General education</p></td><td align="left"><p>77</p></td><td align="left"><p>Broadly related to education, including teacher preparedness, innovative pedagogical frameworks, and AI integration in teaching</p></td></tr><tr><td align="left"><p>Student focus</p></td><td align="left"><p>30</p></td><td align="left"><p>Focus on student interactions with AI, learning experiences, and outcomes in various educational settings</p></td></tr><tr><td align="left"><p>Pedagogy</p></td><td align="left"><p>4</p></td><td align="left"><p>Exploration of pedagogical theories and practices, emphasizing AI support in teaching methodologies</p></td></tr><tr><td align="left"><p>Curriculum design</p></td><td align="left"><p>1</p></td><td align="left"><p>Design and development of curricula, including the role of AI in curriculum planning and implementation</p></td></tr><tr><td align="left"><p>Others</p></td><td align="left"><p>37</p></td><td align="left"><p>Various other contexts not specifically categorized under K-12, higher education, or professional training</p></td></tr></tbody></table> </ephtml> </p> <p>This categorization provides insights into the diverse contexts in which AI is applied in education. While a significant portion of the articles fit into well-defined categories such as K-12, higher education, and professional training, others address broader educational themes that still align with our focus on formal educational settings.</p> <p>We compiled a dataset of 269 selected papers for text mining analysis, including the authors, titles, journal names, publication years, and abstracts of each study. To prepare the abstracts for analysis, we first cleaned the text, removing non-essential characters and standardizing text formats. We then normalized the text by converting it to lowercase and applying stemming and lemmatization, which simplifies words to their root forms. Lastly, we removed stopwords to reduce linguistic noise, ensuring the meaningful patterns in the data were more visible.</p> <hd id="AN0183751070-9">Co-occurrence network</hd> <p>To gain insights into the thematic structure and interrelationships among key concepts within our dataset of educational and scientific abstracts, we performed a co-occurrence network analysis. The first step was to create a co-occurrence matrix by identifying pairs of words that appeared together in the same abstracts, establishing the foundation for our exploration of term interactions. To refine our understanding of these relationships, we applied the <emph>Jaccard index</emph> (Hosseini et al., [<reflink idref="bib36" id="ref59">36</reflink>]; Pourhatami et al., [<reflink idref="bib61" id="ref60">61</reflink>]), defined as:</p> <p> <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>J</mi><mfenced close=")" open="("><mrow><msub><mi>T</mi><mi>i</mi></msub><mo>,</mo><msub><mi>T</mi><mi>j</mi></msub></mrow></mfenced><mo>=</mo><mfrac><msub><mi>N</mi><mrow><mi mathvariant="italic">ij</mi></mrow></msub><mrow><msub><mi>N</mi><mi>i</mi></msub><mo>+</mo><msub><mi>N</mi><mi>j</mi></msub><mo>-</mo><msub><mi>N</mi><mrow><mi mathvariant="italic">ij</mi></mrow></msub></mrow></mfrac></mrow></math> </ephtml> </p> <p>Graph</p> <p>The Jaccard similarity coefficient, <emph>J</emph> ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>T</mi><mi>i</mi></msub><mo>,</mo><msub><mi>T</mi><mi>j</mi></msub></mrow></math> </ephtml> ), was used to quantify the similarity between two terms <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>T</mi><mi>i</mi></msub></math> </ephtml> and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>T</mi><mi>j</mi></msub></math> </ephtml> , indicating the number of abstracts containing both terms, divided by the total number of abstracts where either term appears, corrected for the overlap by subtracting the intersection from the sum of individual occurrences, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>N</mi><mi>i</mi></msub><mo>+</mo><msub><mi>N</mi><mi>j</mi></msub></mrow></math> </ephtml> . This allowed us to focus on the most significant and dominant thematic connections, reducing the noise from less meaningful word pairs (Wu et al., [<reflink idref="bib79" id="ref61">79</reflink>]).</p> <p>We then visualized the interconnected web of terms, using a community detection algorithm to discern clusters of closely related terms. These clusters represent distinct thematic areas in the dataset, revealing the underlying structure of the discourse (Deng et al., [<reflink idref="bib25" id="ref62">25</reflink>]). To further dissect the network's architecture, we calculated two critical measures: centrality and density. <emph>Centrality</emph>, calculated for each node as <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>C</mi><mi>D</mi></msub><mfenced close=")" open="("><mi>t</mi></mfenced><mo>=</mo><mfrac><mrow><mo>deg</mo><mfenced close=")" open="("><mi>t</mi></mfenced></mrow><mrow><mi>N</mi><mo>-</mo><mn>1</mn></mrow></mfrac></mrow></math> </ephtml> , where <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mo>deg</mo><mfenced close=")" open="("><mi>t</mi></mfenced></mrow></math> </ephtml> refer to the connections of the term <emph>t</emph> with other terms and <emph>N</emph> − 1 represents the maximum possible connections that <emph>t</emph> could have with other terms (excluding itself). This measure highlights the most influential words in the dataset. Density, on the other hand, reflects the overall connectivity, calculated using <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>D</mi><mo>=</mo><mfrac><mrow><mn>2</mn><mi>E</mi></mrow><mrow><mi>N</mi><mfenced close=")" open="("><mrow><mi>N</mi><mo>-</mo><mn>1</mn></mrow></mfenced></mrow></mfrac></mrow></math> </ephtml> , where <emph>E</emph> is the number of edges, and <emph>N</emph> is the number of nodes, informing the cohesion of thematic areas.</p> <p>Lastly, we identified the top 30 terms with the highest centrality scores within the co-occurrence network. The identification of key concepts established the stage for our LDA analysis, providing a focused set of terms to explore how thematic patterns are distributed across the corpus.</p> <hd id="AN0183751070-10">Latent Dirichlet allocation</hd> <p>We applied LDA to uncover latent topics within our dataset. LDA, a probabilistic modeling approach, allows for the identification of topics as distributions of words, providing insight into the thematic structures of the corpus without pre-defined categories (Lin et al., [<reflink idref="bib49" id="ref63">49</reflink>]). It assumes that documents (academic abstracts in this study) are created from a mixture of topics, where each word's presence is associated with one of the document's topics (Xing et al., [<reflink idref="bib80" id="ref64">80</reflink>]).</p> <p>Applying LDA to a collection of texts involves analyzing each document to identify its primary themes. Given a corpus containing <emph>D</emph> documents and <emph>K</emph> topics, for each document <emph>d</emph>:</p> <p></p> <ulist> <item> Choose <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>N</mi><mi>d</mi></msub></math> </ephtml> ~ Poisson (<emph>x</emph>), the number of words in document <emph>d</emph>.</item> <p></p> <item> Choose <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>q</mi><mi>d</mi></msub></math> </ephtml> ~ Dirichlet (<emph>a</emph>), the topic distribution for document <emph>d</emph>.</item> <p></p> <item> For each of the <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>N</mi><mi>d</mi></msub></math> </ephtml> words <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>w</mi><mrow><mi>d</mi><mo>,</mo><mi>n</mi></mrow></msub></math> </ephtml> :</item> <p></p> <item> Choose a topic <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>z</mi><mrow><mi>d</mi><mo>,</mo><mi>n</mi></mrow></msub></math> </ephtml> ~ Multinomial ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>q</mi><mi>d</mi></msub></math> </ephtml> )</item> <p></p> <item> Choose a word <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>w</mi><mrow><mi>d</mi><mo>,</mo><mi>n</mi></mrow></msub></math> </ephtml> from <emph>p</emph> ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>w</mi><mrow><mi>d</mi><mo>,</mo><mi>n</mi></mrow></msub></math> </ephtml> | <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>z</mi><mrow><mi>d</mi><mo>,</mo><mi>n</mi></mrow></msub></math> </ephtml> , b), ~ a multinomial probability conditioned on the topic <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>z</mi><mrow><mi>d</mi><mo>,</mo><mi>n</mi></mrow></msub></math> </ephtml> .</item> </ulist> <p>The model estimates the word count of each document and determines the potential topics present. It assigns a topic to each word based on its obtained knowledge. The model then selects a word that aligns with the chosen topic (Blei et al., [<reflink idref="bib13" id="ref65">13</reflink>]). We repeatedly performed this process for all documents, continuously improving our estimations of existing topics and their corresponding words.</p> <p>To ensure topic coherence, we utilized a specialized scoring system. The higher the coherence score, the greater the likelihood that the words within a topic are closely related. After training multiple models and averaging the coherence scores across different runs, we selected the model with the highest average score, representing the most meaningful and interpretable topic distribution. The finalized LDA not only enhances the previous co-occurrence network analysis but as provides a more detailed understanding of the thematic components, enabling informed discussions about educators' adoption of AI. In particular, we paid close attention to how topics related to the challenges of AI integration in education, as highlighted in our literature review, emerged in the analysis.</p> <hd id="AN0183751070-11">Sentiment analysis</hd> <p>Finally, we conducted Sentiment Analysis to capture the academic sentiment towards AI in education within our corpus. Using the Valence Aware Dictionary and sEntiment Reasoner (VADER) Hutto and Gilbert ([<reflink idref="bib40" id="ref66">40</reflink>]) tool, we assessed the emotional valence of the literature. VADER, a lexicon, and rule-based sentiment analysis model, is particularly adept at handling texts from social contexts, including the nuances of slang and emoticons, which are also prevalent in more informal academic communications (Jain et al., [<reflink idref="bib42" id="ref67">42</reflink>]).</p> <p>VADER's analysis calculates sentiment scores using a combination of qualitative and quantitative measures (Reshi et al., [<reflink idref="bib66" id="ref68">66</reflink>]). Each input text is scored based on a summation of the valence scores of each word, adjusted for grammar and syntax rules that influence sentiment. The compound score, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mrow><mtext>Compound Score</mtext></mrow><mo>=</mo><mrow /><msup><mo>∑</mo><mrow><mtext>Valence</mtext><mspace width="0.333333em" /></mrow></msup><mfenced close=")" open="("><mi>w</mi></mfenced><mrow /><mo>×</mo><mrow><mspace width="0.333333em" /><mtext>Heuristics</mtext><mspace width="0.333333em" /></mrow><mfenced close=")" open="("><mi>w</mi></mfenced></mrow></math> </ephtml> , combines <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mrow><mtext>Valence</mtext><mspace width="0.333333em" /></mrow><mfenced close=")" open="("><mi>w</mi></mfenced></mrow></math> </ephtml> , which signifies the pre-determined emotion intensity of word <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mi>w</mi></math> </ephtml> , with <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mrow><mtext>Heuristics</mtext><mspace width="0.333333em" /></mrow><mfenced close=")" open="("><mi>w</mi></mfenced></mrow></math> </ephtml> , which are linguistic rules that modify these intensities based on context.</p> <p>Through this analysis, we processed each abstract to extract sentiment scores, revealing the scholarly perspective towards AI adoption in education. These scores, reflecting a collective emotional stance, highlighted how sentiments within the academic field have changed over time. Given that academic writing often employs an objective tone, we took this into account when interpreting our results, considering both the explicit sentiment expressed and the broader context of academic discourse on AI in education.</p> <p>To further analyze how academic perspectives towards AI in education have evolved over time, we applied a one-way ANOVA to assess the statistical significance of sentiment changes across the years. The ANOVA was chosen to evaluate whether the observed differences in sentiment scores across the years were statistically significant, providing a rigorous assessment of temporal shifts in sentiment. Following the ANOVA, we conducted a post-hoc analysis using Tukey's HSD test. This test was used to identify which specific years showed statistically significant differences in sentiment compared to others, allowing for a more detailed understanding of how sentiments varied over time.</p> <p>During the sentiment analysis, outliers were identified in specific years using the standard criterion of data points falling beyond 1.5 times the interquartile range (IQR) from the first and third quartiles. This method objectively identifies extreme values without bias. To ensure a balanced analysis, these outliers were carefully evaluated for their impact on the results. Rather than altering the original scores, we employed robust statistical techniques, such as median-based measures, which are less affected by outliers. This approach preserved the dataset's integrity, allowing us to accurately reflect general sentiment trends while acknowledging the presence of outliers without letting them skew the overall conclusions.</p> <hd id="AN0183751070-12">Results</hd> <p></p> <hd id="AN0183751070-13">What key concepts emerge in academic discussions about AI integration in educational settings...</hd> <p>The co-occurrence network analysis revealed a group of essential terms that are important to discussion on the use of AI in educational settings. Each node in the network corresponds to a key term, with the size indicating the frequency of the term's occurrence. These edges reflect the co-occurrence of terms within the same abstracts, with thicker lines suggesting more frequent co-occurrences. This network provides a visual representation of the interconnectedness of key concepts in the academic discourse on AI in education.</p> <p>As shown in Fig. 1, terms such as 'ethical,' 'curriculum,' 'AI-powered,' and 'demand' are central to the network. The presence of 'generative' in the network indicates an emphasis on the use of AI to create new content and learning materials.</p> <p>Graph: Fig. 1 Co-occurrence network graph</p> <p>The network also shows 'classroom', 'literacy', and 'pandemic' as significant nodes. Additionally, clusters of terms such as 'decision-making' and 'leadership,' alongside 'AI-powered,' highlight the focus on practical applications of AI in educational management and policymaking.</p> <p>The network analysis revealed several theme groups within the discussion on AI in education. Each cluster, identified through modularity-based community detection, represents a network of closely related terms. While some communities were robust, featuring multiple interconnected terms, others were smaller, consisting of only two terms. The focus on the more substantial clusters, as depicted in Fig. 2, provides a detailed understanding of the dominant themes in the academic conversation.</p> <p>Graph: Fig. 2 Community structures within the network</p> <p>Figure 2a highlights a cluster featuring terms such as 'professional,' 'positive,' 'towards,' 'attitude,' and 'program,' where the central positioning of 'professional' is closely linked to 'positive' and 'program,' with further connections between 'attitude' and terms like 'towards,' 'lack,' 'curriculum,' and 'demand.' In Fig. 2b, the cluster centers around the term 'ethical,' showing connections to 'risk,' 'responsible,' 'clear,' and 'policy,' along with 'generative,' which is associated with 'classroom' and 'science.' Fig. 2c depicts a cluster revolving around the concept of 'role,' establishing links with terms such as 'individual,' 'change,' 'advancement,' and 'highlight.' Meanwhile, Fig. 2d illustrates 'AI-powered' as a core concept connected to terms including 'writing,' 'English as a foreign language (EFL),' 'concern,' 'institution,' and 'privacy.'</p> <p>Table 2 shows the 30 terms with the highest centrality across the entire co-occurrence network. These central terms, ranging from 'decision-making' to 'competency-based', represent key issues including ethical considerations, technological integration, and leadership roles. This quantitative data serves as a foundation for the subsequent analysis of thematic structures through LDA and the examination of evolving academic perspectives through sentiment analysis.</p> <p>Table 2 Co-occurrence matrix</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Term</p></th><th align="left"><p>Centrality</p></th><th align="left"><p>Term</p></th><th align="left"><p>Centrality</p></th><th align="left"><p>Term</p></th><th align="left"><p>Centrality</p></th></tr></thead><tbody><tr><td align="left"><p>Role</p></td><td char="." align="char"><p>0.18666667</p></td><td align="left"><p>Literacy</p></td><td char="." align="char"><p>0.05333333</p></td><td align="left"><p>Writing</p></td><td char="." align="char"><p>0.04</p></td></tr><tr><td align="left"><p>Ethical</p></td><td char="." align="char"><p>0.16</p></td><td align="left"><p>Attitude</p></td><td char="." align="char"><p>0.05333333</p></td><td align="left"><p>Insight</p></td><td char="." align="char"><p>0.04</p></td></tr><tr><td align="left"><p>AI-powered</p></td><td char="." align="char"><p>0.14666667</p></td><td align="left"><p>Machine</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Decision</p></td><td char="." align="char"><p>0.04</p></td></tr><tr><td align="left"><p>Positive</p></td><td char="." align="char"><p>0.13333333</p></td><td align="left"><p>Towards</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Pedagogy</p></td><td char="." align="char"><p>0.02666667</p></td></tr><tr><td align="left"><p>Concern</p></td><td char="." align="char"><p>0.13333333</p></td><td align="left"><p>Growing</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Traditional</p></td><td char="." align="char"><p>0.02666667</p></td></tr><tr><td align="left"><p>ChatGPT</p></td><td char="." align="char"><p>0.12</p></td><td align="left"><p>Program</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Foreign</p></td><td char="." align="char"><p>0.02666667</p></td></tr><tr><td align="left"><p>Professional</p></td><td char="." align="char"><p>0.10666667</p></td><td align="left"><p>EFL</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Individual</p></td><td char="." align="char"><p>0.02666667</p></td></tr><tr><td align="left"><p>Generative</p></td><td char="." align="char"><p>0.08</p></td><td align="left"><p>Society</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Setting</p></td><td char="." align="char"><p>0.02666667</p></td></tr><tr><td align="left"><p>Risk</p></td><td char="." align="char"><p>0.08</p></td><td align="left"><p>Providing</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Change</p></td><td char="." align="char"><p>0.02666667</p></td></tr><tr><td align="left"><p>Classroom</p></td><td char="." align="char"><p>0.06666667</p></td><td align="left"><p>Implementation</p></td><td char="." align="char"><p>0.04</p></td><td align="left"><p>Pandemic</p></td><td char="." align="char"><p>0.02666667</p></td></tr></tbody></table> </ephtml> </p> <p>In summary, the co-occurrence network analysis provides a comprehensive overview of the key concepts and their interconnections in the academic discourse on AI integration in educational settings. This analysis sets the foundation for our exploration of thematic structures through LDA and the evolution of academic perspectives through sentiment analysis, which we will examine in the following sections.</p> <hd id="AN0183751070-14">How do thematic structures identified through latent Dirichlet allocation (LDA) describe the...</hd> <p>Coherence scores was used to assess the model's capability to generate topics that are internally coherent and practically relevant. Coherence scores measure the degree of semantic similarity between high scoring words in each topic. Higher scores indicate more coherent and interpretable topics (Hoyle et al., [<reflink idref="bib37" id="ref69">37</reflink>]; Nanda et al., [<reflink idref="bib54" id="ref70">54</reflink>]). While the models with 2 and 3 topics had higher scores as shown in Fig. 3, we chose a 5-topic model as it provided the best balance between detail and clarity. This decision highlights the significance of thorough testing in determining the optimal number of topics for a model, ensuring precision and comprehensibility (Chu et al., [<reflink idref="bib22" id="ref71">22</reflink>]; Pan & Xue, [<reflink idref="bib58" id="ref72">58</reflink>]).</p> <p>Graph: Fig. 3 Coherence scores by topics</p> <p>Figure 4 presents a map with five distinct topic clusters generated by the LDA algorithm. The map shows these clusters as non-overlapping circles, indicating unique and equally significant topics within our dataset. The placement of the circles reflects the strength of their relationships: close circles suggest closer relation (Egger & Yu, [<reflink idref="bib26" id="ref73">26</reflink>]).</p> <p>Graph: Fig. 4 Intertopic distance map generated by LDA</p> <p>Table 3 summarizes that LDA results by listing the most important terms associated with five topics in our study.</p> <p>Table 3 Five topic model results</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" colspan="2"><p>Topic</p></th><th align="left"><p>Count</p></th><th align="left"><p>Most important terms</p></th></tr></thead><tbody><tr><td align="left"><p>1</p></td><td align="left"><p>Ethical considerations and trust in AI-enhanced literacy education</p></td><td align="left"><p>53</p></td><td align="left"><p>ChatGPT (0.0081), ethical (0.0063), machine* (0.0058), trust* (0.0047), literacy* (0.0040), reading* (0.0039), AI-powered (0.0036), ML* (0.0036), writing (0.0031), AI-EdTech* (0.0029)</p></td></tr><tr><td align="left"><p>2</p></td><td align="left"><p>The role of generative AI in science education and public health</p></td><td align="left"><p>84</p></td><td align="left"><p>ChatGPT (0.0149), writing (0.0115), generative* (0.0058), science (0.0049), instruction* (0.0046), ethical (0.0044), public* (0.0041), AI-powered (0.0039), role (0.0038), health* (0.0037)</p></td></tr><tr><td align="left"><p>3</p></td><td align="left"><p>Leadership and collaborative decision-making in AI-infused classrooms</p></td><td align="left"><p>40</p></td><td align="left"><p>Classroom (0.0058), ethical (0.0050), robot* (0.0049), individual* (0.0047), decision-making* (0.0044), decision* (0.0043), role (0.0040), leadership* (0.0040), collaborative* (0.0037), business* (0.0035)</p></td></tr><tr><td align="left"><p>4</p></td><td align="left"><p>Integrating AI into medical education</p></td><td align="left"><p>58</p></td><td align="left"><p>Medical (0.0059), science (0.0042), legal* (0.0040), intention* (0.0037), attitude* (0.0037), classroom (0.0037), role (0.0034), collaboration* (0.0034), towards* (0.0031), professional (0.0031)</p></td></tr><tr><td align="left"><p>5</p></td><td align="left"><p>Privacy and ethical challenges in AI-powered education</p></td><td align="left"><p>39</p></td><td align="left"><p>ChatGPT (0.0068), professional (0.0052), AI-powered (0.0043), ethical (0.0041), instructor* (0.0041), privacy* (0.0040), algorithm* (0.0036), medical (0.0036), curriculum* (0.0034), pandemic* (0.0029)</p></td></tr></tbody></table> </ephtml> </p> <p>*Words specific to the subject are indicated with an asterisk</p> <p>Topic 1, named as <emph>Ethical Considerations and Trust in AI-enhanced Literacy Education</emph>, centers around keywords such as 'ChatGPT', 'ethical', 'machine', 'trust', 'literacy', 'reading, 'AI-powered', 'ML', 'writing', and 'AI-EdTech'. The emergence of 'machine' and 'trust' in this context may imply the growing importance of transparent, accountable AI systems that educators and learners can rely on. Studies within our dataset, including those by Eguchi et al. ([<reflink idref="bib27" id="ref74">27</reflink>]), Kim and Lee ([<reflink idref="bib46" id="ref75">46</reflink>]), and Relmasira et al. ([<reflink idref="bib65" id="ref76">65</reflink>]), stress the importance of ethical considerations and trust in the adoption of AI technologies in education, promoting the responsible use of AI to enhance literacy skills while compromising ethical values.</p> <p>Topic 2, named as <emph>The Role of Generative AI in Science Education and Public Health</emph>, focuses on terms such as 'ChatGPT', 'writing', 'generative', 'science', 'instruction', 'ethical', 'public', AI-powered', 'role', and 'health.' The distinct emphasis on 'generative', 'public', and 'health' emphasizes its distinctive attention to using generative AI models like ChatGPT in science education and public health contexts. Studies in our dataset indicate that generative AI can significantly contribute to personalized and effective science education, promoting ethical and public health awareness (Akiba & Fraboni, [<reflink idref="bib4" id="ref77">4</reflink>]; Akudjedu et al., [<reflink idref="bib5" id="ref78">5</reflink>]; Wolf & Wolf, [<reflink idref="bib77" id="ref79">77</reflink>]; Wood et al., [<reflink idref="bib78" id="ref80">78</reflink>]).</p> <p>Topic 3, labeled as <emph>Leadership and Collaborative Decision-Making in AI-infused Classrooms</emph>, has 'classroom', 'ethical', 'robot', 'individual', 'decision-making', 'role', 'leadership', 'collaborative', and 'businesses.' The distinct focus on 'robot', 'individual', 'decision-making', 'leadership', and 'collaborative' highlights its concentration on integrating AI to foster leadership and collaborative decision-making in educational settings. Within our dataset, Nguyen et al. ([<reflink idref="bib55" id="ref81">55</reflink>]) investigated the potential use of robots assisting in various university functions, for example, supporting learning enabling remote attendance, and facilitating campus services, therefore encouraging engagement, and reducing faculty workload. Similarly, Ma et al. ([<reflink idref="bib52" id="ref82">52</reflink>]) also explored robots that adapt their engagement based on learning status, enhancing teaching experiences and human-AI interaction. Kajiwara et al. ([<reflink idref="bib43" id="ref83">43</reflink>]) introduced a machine learning role-playing game to clarify AI's educational impact, showing that role-playing may effectively explain AI processes and promote a positive perception of AI among students from elementary levels to the elderly. However, Celik ([<reflink idref="bib17" id="ref84">17</reflink>]) explored the inclusion of ethical considerations in the integration of AI tools, suggesting a delicate approach that balances technological advancement with ethical and thoughtful incorporation into educational practices.</p> <p>Topic 4, <emph>Integrating AI into Medical Education</emph>, focuses on 'medical', 'science', 'intention', 'attitude', 'classroom', 'role', 'collaboration', and 'professional'. This attention to 'intention', 'attitude', and 'collaboration' suggests a dedicated effort to foster multidisciplinary collaborations, addressing the ethical and complex aspects of AI applications in professional education. The literature indicates that such collaborative approaches significantly improved learning outcomes and professional skills in such specialized domains (Abdellatif et al., [<reflink idref="bib1" id="ref85">1</reflink>]; Alghamdi & Alashban, [<reflink idref="bib7" id="ref86">7</reflink>]; Truong et al., [<reflink idref="bib71" id="ref87">71</reflink>]).</p> <p>Lastly, Topic 5, <emph>Privacy and Ethical Challenges in AI-powered Education</emph>, has the unique emphasis on 'AI-powered', 'instructor', 'privacy', 'algorithm', 'curriculum', and 'pandemic'. The specific emphasis on 'curriculum' reflects the need for designing AI-integrated curricula that meet the changing educational demands (Wang, [<reflink idref="bib74" id="ref88">74</reflink>]). Another prevalent term 'algorithm' indicates the critical evaluation of AI processes for fairness and bias mitigation (Velander et al., [<reflink idref="bib72" id="ref89">72</reflink>]). The term 'pandemic' demonstrates the increased use of AI technologies during global health crises.</p> <p>These identified themes both align with and extend previous research on AI in education. For instance, the emphasis on ethical considerations across multiple topics echoes the findings of Zawacki-Richter et al. ([<reflink idref="bib81" id="ref90">81</reflink>]), who highlighted the growing importance of ethics in AI-enhanced education. However, our analysis reveals a deeper understanding of these ethical concerns, particularly in relation to specific contexts such as literacy education and medical training. The emergence of generative AI as a distinct topic represents a shift from earlier studies, reflecting the rapid advancement of these technologies and their increasing relevance in educational settings.</p> <hd id="AN0183751070-15">How has the academic perspective towards AI in education evolved over time, as reflected in s...</hd> <p>Figure 5 illustrates the evolution of academic and emotional perspectives toward the use of AI by educators, as inferred from sentiment analysis of relevant publications from 2018 to 2024. The sentiment scores, ranging from − 1 (entirely negative) to + 1 (entirely positive), continue to reveal a generally positive median sentiment across the years. However, certain years exhibit distinct variations in sentiment, reflecting changing attitudes toward AI's role in education.</p> <p>Graph: Fig. 5 Sentiment score distribution by year</p> <p>To assess the statistical significance of these sentiment changes over time, a one-way ANOVA test was conducted, which revealed a significant difference in sentiment scores across years [F(<reflink idref="bib6" id="ref91">6</reflink>, 267) = 5.35, p <.01]. This result confirms that the observed changes in academic perspectives represent meaningful shifts in sentiment over time, rather than random variation.</p> <p>Year 2020 stands out with a notably wider interquartile range and a lower median sentiment, indicating a more varied and somewhat negative response from educators during the COVID-19 pandemic. This variation likely reflects the challenges associated with the rapid adoption of AI tools during this period, where educators had to quickly adapt to AI-supported remote teaching.</p> <p>In contrast, the year 2023 presents a slightly wider interquartile range with a lower median sentiment, indicating a more negative and diverse set of reactions toward AI tools. The distribution of sentiment scores in 2023 suggests ongoing concerns related to the integration of AI technologies, such as ChatGPT, into educational settings. Publications from 2023, including those by Crawford et al. ([<reflink idref="bib23" id="ref92">23</reflink>]), Hung and Chen ([<reflink idref="bib39" id="ref93">39</reflink>]), Popenici ([<reflink idref="bib60" id="ref94">60</reflink>]), Velander et al. ([<reflink idref="bib72" id="ref95">72</reflink>]), and Zou and Huang ([<reflink idref="bib83" id="ref96">83</reflink>]), delve into critical debates surrounding the practical and ethical implications of AI in teaching. These diverse sentiment scores highlight the challenges educators face in ensuring responsible AI implementation in education.</p> <p>Year 2024 shows a recovery with a higher median sentiment and less variability, suggesting a more stable and positive outlook on the use of AI in education. This stabilization could reflect a growing familiarity and comfort with AI tools among educators, although some concerns remain, as evidenced by the range of sentiment scores.</p> <p>As shown in Table 4, the post-hoc analysis, using Tukey's HSD test, further explained the specific differences between years. The sentiment score in 2024 was significantly higher than in 2020 (p =.0147), indicating a recovery in sentiment following the challenges of the earlier pandemic period. Additionally, a significant difference was observed between 2022 and 2023 (p =.0164), with 2023 exhibiting more negative sentiment, reflecting concerns as AI tools like ChatGPT became more integrated into educational settings. The contrast between 2023 and 2024 also displayed a significant difference (p =.0002), underscoring the shift in sentiment as the academic community continued to adapt to AI's evolving role in education.</p> <p>Table 4 Post-hoc Tukey HSD results</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Group 1</p></th><th align="left"><p>Group 2</p></th><th align="left"><p>Mean difference</p></th><th align="left"><p>p-value</p></th><th align="left"><p>Reject <math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>H</mi><mn>0</mn></msub></math><inline-graphic mime-subtype="GIF" href="11423_2024_10425_Article_IEq22.gif" /></p></th></tr></thead><tbody><tr><td align="left" rowspan="6"><p>2018</p></td><td align="left"><p>2019</p></td><td char="." align="char"><p>.0756</p></td><td char="." align="char"><p>1</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2020</p></td><td char="." align="char"><p> −.358</p></td><td char="." align="char"><p>.9487</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2021</p></td><td char="." align="char"><p>.0689</p></td><td char="." align="char"><p>1</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2022</p></td><td char="." align="char"><p>.0699</p></td><td char="." align="char"><p>1</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2023</p></td><td char="." align="char"><p> −.2525</p></td><td char="." align="char"><p>.9827</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2024</p></td><td char="." align="char"><p>.3771</p></td><td char="." align="char"><p>.9141</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left" rowspan="5"><p>2019</p></td><td align="left"><p>2020</p></td><td char="." align="char"><p> −.4336</p></td><td char="." align="char"><p>.7047</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2021</p></td><td char="." align="char"><p> −.0067</p></td><td char="." align="char"><p>1</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2022</p></td><td char="." align="char"><p> −.0057</p></td><td char="." align="char"><p>1</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2023</p></td><td char="." align="char"><p> −.3281</p></td><td char="." align="char"><p>.7511</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2024</p></td><td char="." align="char"><p>.3015</p></td><td char="." align="char"><p>.8926</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left" rowspan="4"><p><bold>2020</bold></p></td><td align="left"><p>2021</p></td><td char="." align="char"><p>.4269</p></td><td char="." align="char"><p>.4075</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2022</p></td><td char="." align="char"><p>.4279</p></td><td char="." align="char"><p>.2943</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2023</p></td><td char="." align="char"><p>.1055</p></td><td char="." align="char"><p>.9974</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p><bold>2024</bold></p></td><td char="." align="char"><p><bold>.7351</bold></p></td><td char="." align="char"><p><bold>.0147*</bold></p></td><td align="left"><p><bold>TRUE</bold></p></td></tr><tr><td align="left" rowspan="3"><p>2021</p></td><td align="left"><p>2022</p></td><td char="." align="char"><p>.001</p></td><td char="." align="char"><p>1</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2023</p></td><td char="." align="char"><p> −.3214</p></td><td char="." align="char"><p>.1671</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p>2024</p></td><td char="." align="char"><p>.3082</p></td><td char="." align="char"><p>.5811</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left" rowspan="2"><p><bold>2022</bold></p></td><td align="left"><p><bold>2023</bold></p></td><td char="." align="char"><p> − <bold>.3224</bold></p></td><td char="." align="char"><p><bold>.0164*</bold></p></td><td align="left"><p><bold>TRUE</bold></p></td></tr><tr><td align="left"><p>2024</p></td><td char="." align="char"><p>.3072</p></td><td char="." align="char"><p>.4164</p></td><td align="left"><p>FALSE</p></td></tr><tr><td align="left"><p><bold>2023</bold></p></td><td align="left"><p><bold>2024</bold></p></td><td char="." align="char"><p><bold>.6296</bold></p></td><td char="." align="char"><p><bold>.0002**</bold></p></td><td align="left"><p><bold>TRUE</bold></p></td></tr></tbody></table> </ephtml> </p> <p>*p < 0.05, **p < 0.01</p> <p>Overall, the sentiment analysis indicates a predominantly positive sentiment towards AI in education throughout the years analyzed, with some variability reflecting specific contextual factors.</p> <hd id="AN0183751070-16">Discussion</hd> <p>This study aimed to provide a comprehensive understanding of the academic discourse surrounding AI integration in educational settings by investigating three key research questions. The findings from the co-occurrence network analysis, LDA, and sentiment analysis offer valuable insights into the key concepts, thematic structures, and evolving perspectives on AI integration in educational settings.</p> <hd id="AN0183751070-17">Ethical considerations and global events impacting AI integration</hd> <p>The co-occurrence network analysis revealed the centrality of ethical considerations in the academic discourse on AI integration in education. The prominence of terms like 'ethical,' 'AI-powered,' and 'curriculum' emphasizes the strong focus on the ethical implications of integrating AI into educational practices. This centrality of ethics suggests a deep concern within the academic community about ensuring that AI's incorporation into education is aligned with moral and ethical standards. Such focus aligns with the findings of Berendt et al. ([<reflink idref="bib11" id="ref97">11</reflink>]) and other studies that highlight the need for careful consideration of ethical issues in AI adoption in education (Adams et al., [<reflink idref="bib2" id="ref98">2</reflink>]; Casal-Otero et al., [<reflink idref="bib16" id="ref99">16</reflink>]; Celik et al., [<reflink idref="bib18" id="ref100">18</reflink>]; Qin et al., [<reflink idref="bib63" id="ref101">63</reflink>]).</p> <p>Moreover, the analysis revealed a connection between 'pandemic' and terms like 'remote' and 'AI-powered,' indicating how the COVID-19 pandemic acted as a catalyst for the rapid adoption of AI tools, particularly in facilitating remote learning (see Fig. 1). This connection highlights the dual role of AI in providing innovative solutions for unprecedented challenges and raising new ethical considerations in its deployment during crises. The increased reliance on AI during the pandemic may have accelerated discussions about the ethical implications of such technologies, as educators and institutions had to adapt quickly to maintain continuity in learning.</p> <p>Additionally, the identification of terms like 'positive,' 'towards,' and 'program' in the network (see Fig. 2a) suggests a significant discussion within the academic community about the generally favorable attitudes of educational professionals towards AI programs. This growing acceptance and eagerness to adopt AI within educational practices reflect a broader trend of increasing confidence in AI's potential to enhance educational outcomes. However, this trend also brings to light the need for ongoing scrutiny of AI's impact, especially regarding how these tools influence curriculum design and the educational experience.</p> <p>The term 'lack' in the network analysis (see Fig. 2a) might point to areas where AI could fill gaps in the curriculum or where improvements in AI-related educational content are needed. This highlights an ongoing concern about ensuring that AI not only supplements but also enhances the educational curriculum without creating dependencies that might undermine the integrity of educational practices.</p> <p>The network also indicated a central concern for ethical considerations in the use of chatbots and generative AI tools like ChatGPT, particularly in shaping literacy education. The connection of 'literacy' with ethical considerations suggests that while AI offers innovative methods to improve literacy, it also necessitates careful management to ensure these tools do not compromise educational quality or ethical standards.</p> <p>Further, the presence of terms like 'innovation' and 'comprehensive' in the network (see Fig. 2b) points to a broader perspective on AI's incorporation into educational practices, reflecting a recognition that AI's role in education is expansive and multi-faceted. This perspective is essential as it suggests that while AI offers opportunities for innovation, it also requires a comprehensive approach to implementation—one that considers the full spectrum of educational, ethical, and societal implications.</p> <p>Additionally, the analysis showed that the discussion around AI integration in education often contrasts traditional methods with machine-assisted techniques (see Fig. 2c). This contrast underlines a broader dialogue about the evolving role of technology in education and how traditional practices are being challenged or augmented by AI. The term 'traditional' linked with 'machine' suggests a transition phase where educators and institutions are negotiating the balance between maintaining proven educational practices and embracing AI's potential to innovate.</p> <p>The proximity of the term 'foreign' with AI-powered tools (see Fig. 2d) indicates a particular emphasis on AI's role in foreign language education. This focus suggests that AI is being seen as a valuable tool in overcoming linguistic barriers and enhancing language learning, but it also raises questions about the implications of relying heavily on AI for such personalized aspects of education.</p> <p>Finally, the highlighted privacy implications of using AI technologies in educational settings, especially in the context of data protection and student confidentiality (see Fig. 2d), reinforce the need for robust ethical frameworks. These frameworks must address the dual challenge of leveraging AI's capabilities while safeguarding the privacy and integrity of educational environments. The connection between these ethical considerations and emerging global events suggests that the academic discourse on AI in education is not static but responsive to broader societal changes and challenges, indicating a dynamic and evolving field of study.</p> <hd id="AN0183751070-18">Thematic structures and privacy challenges in AI-powered education</hd> <p>The LDA analysis provided deeper insights into how AI is integrated into educational practices, revealing distinct thematic structures that feature the multifaceted role of AI in education. The clear distinction among the topics identified by the LDA model (see Fig. 4) highlights the unique importance of each thematic area, reflecting the model's effectiveness in maintaining their relevance within the broader discourse on AI in education.</p> <hd id="AN0183751070-19">Topic 1: Ethical considerations and trust in AI-enhanced literacy education</hd> <p>One of the prominent themes identified was the ethical utilization of AI and machine learning (ML) in literacy education. The LDA analysis highlighted terms such as 'machine,' 'trust,' 'literacy,' 'reading,' and 'AI-EdTech,' indicating a critical focus on ensuring that AI enhances literacy skills while maintaining ethical standards and trust in technology. This emphasis on ethics reflects an integrated approach to teaching, where the effectiveness of AI tools is balanced with the responsibility to uphold ethical values. Such concerns are particularly important in literacy education, where AI's direct interaction with learners necessitates careful consideration of its ethical implications. The focus on trust and ethics in this context suggests that while AI offers innovative solutions, it must be implemented in a way that preserves the integrity of the educational experience.</p> <hd id="AN0183751070-20">Topic 2: The role of generative AI in science education and public health</hd> <p>The analysis also brought to light the transformative potential of generative AI in science education and public health. This topic illustrates how AI can enhance educational content creation and address health challenges through innovative instructional strategies. The emphasis on terms like 'ChatGPT,' 'writing,' 'generative,' and 'science' points to AI's capacity to significantly improve the quality of instructional materials, thereby advancing educational outcomes in specialized fields where accuracy and precision are critical. The focus on public health further demonstrates AI's broader societal impact, highlighting its potential to contribute to fields that extend beyond traditional educational boundaries. These transformative potential stresses the need for careful integration of AI technologies to maximize their benefits while managing associated risks.</p> <hd id="AN0183751070-21">Topic 3: Leadership and collaborative decision-making in AI-infused classrooms</hd> <p>Another significant theme identified was the role of AI in fostering leadership and collaborative decision-making in educational settings. The terms 'leadership,' 'decision-making,' and 'collaborative' indicate a growing interest in how AI can support these critical aspects of educational management. The LDA analysis suggests that AI's role in education extends beyond instructional support to include the enhancement of leadership and decision-making processes, particularly through the innovative use of AI-driven tools such as robots in classrooms. This theme highlights the evolving nature of educational innovation, where AI not only aids in teaching but also plays a pivotal role in shaping how educational institutions are managed and how decisions are made collaboratively.</p> <hd id="AN0183751070-22">Topic 4: Integrating AI into medical education</hd> <p>The integration of AI into medical education emerged as another key thematic structure. The terms 'medical,' 'collaboration,' and 'intention' reflect the focused effort to incorporate AI tools into the professional training of medical practitioners. This theme stresses the importance of AI in enhancing learning outcomes in highly specialized and complex domains such as medicine. The emphasis on collaboration also points to the multidisciplinary nature of AI's application in medical education, where the convergence of various fields can lead to more effective and innovative training methods. This integration of AI into medical education aligns with broader trends in the adoption of AI technologies in professional fields that require precise and rigorous training.</p> <hd id="AN0183751070-23">Topic 5: Privacy and ethical challenges in AI-powered education</hd> <p>Privacy and ethical challenges associated with AI-powered education were highlighted as significant concerns in the LDA analysis. The emphasis on terms like 'privacy,' 'algorithm,' and 'pandemic' reveals ongoing discussions about the protection of student and instructor data in digital learning environments. As AI technologies increasingly collect and analyze vast amounts of personal data, the risks associated with data security, consent, and algorithmic bias become more pronounced. This theme highlights the necessity for robust ethical frameworks and professional development programs that equip educators with the skills to navigate these challenges effectively. The analysis also suggests that the pandemic has exacerbated these concerns, as the rapid shift to remote learning has accelerated the adoption of AI technologies, often without sufficient consideration of privacy implications.</p> <hd id="AN0183751070-24">Interconnections and implications</hd> <p>The interconnections between these topics are particularly significant. For example, the ethical considerations highlighted in AI-enhanced literacy education (Topic 1) are likely to influence the integration of AI in medical education (Topic 4) and the broader challenges of privacy in AI-powered education (Topic 5). Similarly, the focus on leadership and collaborative decision-making (Topic 3) could inform how generative AI is implemented in science education (Topic 2), suggesting a cross-pollination of ideas and concerns across different educational contexts.</p> <p>These findings have significant implications for stakeholders in education. Educators need ongoing professional development to navigate the ethical challenges of AI integration while maximizing its potential in areas like literacy and science education. Policymakers should consider developing comprehensive guidelines that address privacy concerns and the ethical use of AI across different educational contexts, with an emphasis on adaptability to the rapid advancements in AI technology. Researchers are encouraged to further investigate the long-term impacts of AI integration on learning outcomes and pedagogical practices, as the thematic structures identified suggest promising areas for future exploration.</p> <p>Finally, it is important to acknowledge the limitations of LDA in this analysis. While LDA provides valuable insights into the thematic structures of academic discourse, it assumes that each document is a mixture of topics, which may not always capture the full complexity of the discourse. Additionally, the interpretation of topics can be subjective, and there may be overlap between themes that is not fully captured in this analysis (Chen et al., [<reflink idref="bib19" id="ref102">19</reflink>]). Despite these limitations, the findings offer a robust foundation for understanding the current landscape of AI integration in education and highlight critical areas that require ongoing attention and research.</p> <hd id="AN0183751070-25">Evolving sentiments and challenges over time</hd> <p>The sentiment analysis offered a temporal perspective on the evolving academic attitudes toward AI in education. While the overall sentiment remained predominantly optimistic, reflecting the growing interest in using AI to improve educational practices (Al Darayseh, [<reflink idref="bib6" id="ref103">6</reflink>]; Chiu et al., [<reflink idref="bib21" id="ref104">21</reflink>]), specific periods, such as 2020 and 2023, showed notable dips in sentiment.</p> <p>The year 2020, marked by the rapid adoption of AI-driven remote learning tools due to the COVID-19 pandemic, revealed mixed sentiments. Educators were faced with the dual challenge of leveraging AI's potential to maintain educational continuity while coping with the sudden transition to online learning. The lower median sentiment observed during this period likely reflects the stress and uncertainty experienced by educators, as they recognized AI's potential but also struggled with its immediate implementation under unprecedented circumstances. This mix of reactions highlights the complexity of AI integration during a crisis, where the need for technological solutions was urgent, but the infrastructure and support for such a shift were not fully in place.</p> <p>In 2023, the emergence of negative outliers in sentiment coincided with the rise of advanced AI systems like ChatGPT. This period saw increasing concerns about AI's impact on academic integrity, particularly regarding the changing nature of assessment. The critical perspectives emerging during this time suggest that while AI tools like ChatGPT offered innovative educational possibilities, they also raised significant ethical questions. These concerns included the potential misuse of AI technologies, the effectiveness of teaching when heavily reliant on AI, and the broader implications for academic standards. This critical point in the discourse underscores the need for ongoing dialogue and examination of AI's role in education, as the rapid development of these technologies continues to challenge existing educational paradigms.</p> <p>Despite these challenges, the overarching evidence from the sentiment analysis indicates that the academic community has maintained a predominantly optimistic view of AI's role in education over the years. This positive trend aligns with the increasing interest in harnessing AI to improve educational practices, reflecting a broader recognition of AI's potential to enhance learning outcomes, streamline educational processes, and address diverse educational needs. However, this optimism is tempered by the complexities of AI integration, which requires balancing the potential benefits of AI with ethical considerations and the challenges of effective implementation.</p> <p>These sentiment trends in academic literature also mirror broader trends in AI adoption and public perception during the same periods. For instance, the mixed sentiments observed in 2020 are consistent with the general public's ambivalence toward increased reliance on technology during the pandemic (Araujo et al., [<reflink idref="bib10" id="ref105">10</reflink>]; Glikson & Woolley, [<reflink idref="bib33" id="ref106">33</reflink>]). The surge of critical perspectives in 2023 parallels wider societal debates about the implications of advanced AI systems like ChatGPT, where concerns about privacy, bias, and the erosion of traditional educational values became more pronounced (Lund & Wang, [<reflink idref="bib51" id="ref107">51</reflink>]; Rahman & Watanobe, [<reflink idref="bib64" id="ref108">64</reflink>]).</p> <p>However, it is important to acknowledge the limitations of sentiment analysis when applied to academic writing. Academic papers often maintain a neutral tone, which can result in more muted sentiment scores compared to other forms of discourse. Additionally, sentiment analysis may not fully capture the complexity of arguments that present both positive and negative aspects of AI in education, leading to potential oversimplification of nuanced perspectives.</p> <p>Beyond the COVID-19 pandemic, other contextual factors likely influenced these sentiment trends. The rapid advancement of AI technologies, particularly in natural language processing, has continually reshaped the context of AI in education (Alqahtani et al., [<reflink idref="bib9" id="ref109">9</reflink>]). Policy changes, such as the implementation of AI ethics guidelines in various countries, have also played a role in shaping academic perspectives during this period (Khowaja et al., [<reflink idref="bib45" id="ref110">45</reflink>]). These factors highlight the dynamic nature of AI's integration into education and the need for ongoing research to monitor and understand these evolving sentiments.</p> <hd id="AN0183751070-26">Implications for stakeholders</hd> <p>The findings of this study have significant implications for various stakeholders in education. For educators, there is a clear need to develop comprehensive AI literacy programs that extend beyond technical skills to include ethical considerations and critical evaluation of AI tools. This should involve workshops on responsible AI use in the classroom, strategies for maintaining academic integrity in the age of generative AI, and professional development programs that address privacy and ethical concerns.</p> <p>For policymakers, the study emphasizes the importance of establishing clear guidelines for AI use in educational settings. These guidelines should address data privacy, algorithmic bias, and the ethical use of AI in assessment. Given the rapid pace of technological advancement, it is essential that these guidelines remain flexible and adaptable to future developments.</p> <p>Educational institutions should consider investing in infrastructure and professional development programs that support effective AI integration. This may involve creating AI ethics committees to oversee the implementation of AI technologies and providing ongoing support for educators as they navigate new AI tools.</p> <p>Furthermore, the study highlights the importance of close collaboration between EdTech developers, educators, and researchers in developing AI tools that align with pedagogical needs and ethical standards. Developers should focus on creating transparent AI systems and incorporating features that support, rather than replace, critical thinking skills in students.</p> <hd id="AN0183751070-27">Limitations of the study</hd> <p>Despite the valuable insights provided by this study, it is important to acknowledge its limitations. First, the findings are based on a specific dataset of academic publications, which may not fully represent the perspectives of all educators and stakeholders involved in AI integration in education. Future research could explore a broader range of data sources, such as surveys, interviews, and case studies, to capture a more comprehensive understanding of teacher-AI collaboration.</p> <p>Second, the rapid pace of technological advancements and the evolving nature of AI in education may limit the generalizability of the findings over time. As AI technologies continue to progress and new tools and applications emerge, the dynamics of teacher-AI collaboration may shift. Therefore, ongoing research is necessary to keep pace with these developments and provide up-to-date insights into the challenges and opportunities regarding AI integration in education.</p> <p>Third, while this study provides a comprehensive analysis of the key concepts, thematic structures, and evolving perspectives on AI integration in education, the diversity of educational contexts and AI applications means that some specific insights may not be fully generalizable. Future research could focus on identifying and evaluating the most successful approaches with considerations of pedagogical frameworks, professional development programs, and technological infrastructures.</p> <p>Moreover, future research could explore the long-term impact of teacher-AI collaboration on student learning outcomes, focusing on both cognitive and non-cognitive factors. Additionally, the development and validation of ethical guidelines and practices for responsible and equitable teacher-AI collaboration is another critical area for future research.</p> <hd id="AN0183751070-28">Conclusion</hd> <p>Our study provides novel insights into the academic discourse surrounding AI integration in educational settings, revealing key concepts, thematic structures, and sentiment trends. The co-occurrence network analysis revealed the centrality of ethical considerations and the impact of global events on AI adoption. Our LDA analysis uncovered diverse themes, from the role of generative AI in science education to privacy challenges in AI-powered education. The sentiment analysis demonstrated a generally positive trend towards AI in education, with notable fluctuations during significant events like the COVID-19 pandemic. For educators, policymakers, and EdTech developers, our results emphasize the need for comprehensive AI literacy programs, flexible ethical guidelines, and pedagogically sound tools.</p> <p>As we move forward in this rapidly evolving field, future research should explore the long-term impacts of AI integration on learning outcomes and develop robust ethical frameworks. In embracing the AI revolution in education, we must remain committed to fostering responsible teacher-AI collaboration, addressing ethical and privacy concerns while leveraging AI's potential to enhance learning experiences. By doing so, we can shape a future where AI and human intelligence are seamlessly integrated, fostering innovation and excellence in education, while safeguarding the irreplaceable value of human interaction in the learning process.</p> <hd id="AN0183751070-29">Data availability</hd> <p>The dataset analyzed in this study is available from the corresponding author upon reasonable request.</p> <hd id="AN0183751070-30">Declarations</hd> <p></p> <hd id="AN0183751070-31">Conflicts of interest</hd> <p>The authors declare that they have no conflicts of interest.</p> <hd id="AN0183751070-32">Publisher's Note</hd> <p>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p> <ref id="AN0183751070-33"> <title> References </title> <blist> <bibl id="bib1" idref="ref52" type="bt">1</bibl> <bibtext> Abdellatif H, Al Mushaiqri M, Albalushi H, Al-Zaabi AA, Roychoudhury S, Das S. Teaching, learning and assessing anatomy with artificial intelligence: The road to a better future. International Journal of Environmental Research and Public Health. 2022; 19; 21: 14209</bibtext> </blist> <blist> <bibl id="bib2" idref="ref53" type="bt">2</bibl> <bibtext> Adams C, Pente P, Lemermeyer G, Rockwell G. Ethical principles for artificial intelligence in K-12 education. Computers and Education. 2023; 4: 100131</bibtext> </blist> <blist> <bibl id="bib3" idref="ref25" type="bt">3</bibl> <bibtext> Akgun S, Greenhow C. Artificial intelligence in education: Addressing ethical challenges in K-12 settings. AI and Ethics. 2022; 2; 3: 431-440</bibtext> </blist> <blist> <bibl id="bib4" idref="ref55" type="bt">4</bibl> <bibtext> Akiba D, Fraboni MC. AI-supported academic advising: Exploring ChatGPT's current state and future potential toward student empowerment. Education Sciences. 2023; 13; 9: 885. 10.3390/educsci13090885</bibtext> </blist> <blist> <bibl id="bib5" idref="ref56" type="bt">5</bibl> <bibtext> Akudjedu TN, Torre S, Khine R, Katsifarakis D, Newman D, Malamateniou C. Knowledge, perceptions, and expectations of Artificial intelligence in radiography practice: A global radiography workforce survey. Journal of Medical Imaging and Radiation Sciences. 2023; 54; 1: 104-116</bibtext> </blist> <blist> <bibl id="bib6" idref="ref57" type="bt">6</bibl> <bibtext> Al Darayseh A. Acceptance of artificial intelligence in teaching science: Science teachers' perspective. Computers and Education. 2023; 4: 100132</bibtext> </blist> <blist> <bibl id="bib7" idref="ref86" type="bt">7</bibl> <bibtext> Alghamdi SA, Alashban Y. Knowledge, attitudes and practices towards artificial intelligence (AI) among radiologists in Saudi Arabia. Journal of Radiation Research and Applied Sciences. 2023; 16; 2: 100569</bibtext> </blist> <blist> <bibl id="bib8" idref="ref3" type="bt">8</bibl> <bibtext> Alkhatlan, A, & Kalita, J. (2018). Intelligent tutoring systems: A comprehensive historical survey with recent developments. Retrieved from https://arxiv.org/abs/1812.09628</bibtext> </blist> <blist> <bibl id="bib9" idref="ref109" type="bt">9</bibl> <bibtext> Alqahtani T, Badreldin HA, Alrashed M, Alshaya AI, Alghamdi SS, Bin Saleh K, Alowais SA, Alshaya OA, Rahman I, Al Yami MS, Albekairy AM. The emergent role of artificial intelligence, natural learning processing, and large language models in higher education and research. Research in Social and Administrative Pharmacy. 2023; 19; 8: 1236-1242</bibtext> </blist> <blist> <bibtext> Araujo T, Helberger N, Kruikemeier S, De Vreese CH. In AI we trust? Perceptions about automated decision-making by artificial intelligence. AI & Society. 2020; 35; 3: 611-623</bibtext> </blist> <blist> <bibtext> Berendt B, Littlejohn A, Blakemore M. AI in education: Learner choice and fundamental rights. Learning, Media and Technology. 2020; 45; 3: 312-324. 10.1080/17439884.2020.1786399</bibtext> </blist> <blist> <bibtext> Bhutoria A. Personalized education and artificial intelligence in the United States, China, and India: A systematic review using a human-in-the-loop model. Computers and Education. 2022; 3: 100068. 10.1016/j.caeai.2022.100068</bibtext> </blist> <blist> <bibtext> Blei DM, Ng AY, Jordan MI. Latent Dirichlet allocation. Journal of Machine Learning Research. 2003; 3: 993-1022</bibtext> </blist> <blist> <bibtext> Bower M, Torrington J, Lai JW, Petocz P, Alfano M. How should we change teaching and assessment in response to increasingly powerful generative Artificial Intelligence? Outcomes of the ChatGPT teacher survey. Education and Information Technologies. 2024; 29: 1-37</bibtext> </blist> <blist> <bibtext> Božić V. Artificial intelligence as the reason and the solution of digital divide. Language Education and Technology. 2023; 3; 2: 1</bibtext> </blist> <blist> <bibtext> Casal-Otero L, Catala A, Fernández-Morante C, Taboada M, Cebreiro B, Barro S. AI literacy in K-12: A systematic literature review. International Journal of STEM Education. 2023; 10; 1: 29. 10.1186/s40594-023-00418-7</bibtext> </blist> <blist> <bibtext> Celik I. Towards intelligent-TPACK: An empirical study on teachers' professional knowledge to ethically integrate artificial intelligence (AI)-based tools into education. Computers in Human Behavior. 2023; 138: 107468</bibtext> </blist> <blist> <bibtext> Celik I, Dindar M, Muukkonen H, Järvelä S. The promises and challenges of artificial intelligence for teachers: A systematic review of research. TechTrends. 2022; 66; 4: 616-630. 10.1007/s11528-022-00715-y</bibtext> </blist> <blist> <bibtext> Chen X, Zou D, Xie H. A decade of learning analytics: Structural topic modeling based bibliometric analysis. Education and Information Technologies. 2022; 27; 8: 10517-10561. 10.1007/s10639-022-11046-z</bibtext> </blist> <blist> <bibtext> Cheng L, Umapathy K, Rehman M, Ritzhaupt A, Antonyan K, Shidfar P, Nichols J, Lee M, Abramowitz B. Designing, developing, and validating a measure of undergraduate students' conceptions of artificial intelligence in education. Journal of Interactive Learning Research. 2023; 34; 2: 275-311</bibtext> </blist> <blist> <bibtext> Chiu TKF, Xia Q, Zhou X, Chai CS, Cheng M. Systematic literature review on opportunities, challenges, and future research recommendations of artificial intelligence in education. Computers and Education. 2023; 4: 100118. 10.1016/j.caeai.2022.100118</bibtext> </blist> <blist> <bibtext> Chu CY, Park K, Kremer GE. A global supply chain risk management framework: An application of text-mining to identify region-specific supply chain risks. Advanced Engineering Informatics. 2020; 45: 101053</bibtext> </blist> <blist> <bibtext> Crawford J, Cowling M, Allen KA. Leadership is needed for ethical ChatGPT: Character, assessment, and learning using artificial intelligence (AI). Journal of University Teaching & Learning Practice. 2023; 20; 3: 2</bibtext> </blist> <blist> <bibtext> Crompton H, Burke D. Artificial Intelligence in higher education: The state of the field. International Journal of Educational Technology in Higher Education. 2023. 10.1186/s41239-023-00392-8</bibtext> </blist> <blist> <bibtext> Deng S, Xia S, Hu J, Li H, Liu Y. Exploring the topic structure and evolution of associations in information behavior research through co-word analysis. Journal of Librarianship and Information Science. 2021; 53; 2: 280-297. 10.1177/09610006209381</bibtext> </blist> <blist> <bibtext> Egger R, Yu J. A topic modeling comparison between lda, nmf, top2vec, and bertopic to demystify twitter posts. Frontiers in Sociology. 2022; 7: 886498. 10.3389/fsoc.2022.886498</bibtext> </blist> <blist> <bibtext> Eguchi A, Okada H, Muto Y. Contextualizing AI education for K-12 students to enhance their learning of AI literacy through culturally responsive approaches. KI—Künstliche Intelligenz. 2021; 35; 2: 153-161</bibtext> </blist> <blist> <bibtext> Ehimuan B, Anyanwu A, Olorunsogo T, Akindote OJ, Abrahams TO, Reis O. Digital inclusion initiatives: Bridging the connectivity gap in Africa and the USA—A review. International Journal of Science and Research Archive. 2024; 11; 1: 488-501</bibtext> </blist> <blist> <bibtext> Ertmer PA, Ottenbreit-Leftwich AT, Sadik O, Sendurur E, Sendurur P. Teacher beliefs and technology integration practices: A critical relationship. Computers & Education. 2012; 59; 2: 423-435</bibtext> </blist> <blist> <bibtext> Farrokhnia M, Banihashem SK, Noroozi O, Wals A. A SWOT analysis of CHATGPT: Implications for educational practice and research. Innovations in Education and Teaching International. 2023. 10.1080/14703297.2023.2195846</bibtext> </blist> <blist> <bibtext> Ghamrawi N, Shal T, Ghamrawi NA. Exploring the impact of AI on teacher leadership: Regressing or expanding?. Education and Information Technologies. 2023; 7: 1-19</bibtext> </blist> <blist> <bibtext> Gillani, N, Eynon, R, Chiabaut, C, & Finkel, K. (2023). Unpacking the "Black Box" of AI in education. Educational Technology & Society,26(1), 99–111. https://<ulink href="http://www.jstor.org/stable/48707970">www.jstor.org/stable/48707970</ulink></bibtext> </blist> <blist> <bibtext> Glikson E, Woolley AW. Human trust in artificial intelligence: Review of empirical research. Academy of Management Annals. 2020; 14; 2: 627-660</bibtext> </blist> <blist> <bibtext> Hellas, A, Leinonen, J, & Ihantola, P. (2017). Plagiarism in take-home exams: help-seeking, collaboration, and systematic cheating. In: Proceedings of the 2017 ACM Conference on Innovation and Technology in Computer Science Education (pp. 238–243).</bibtext> </blist> <blist> <bibtext> Holmes W, Tuomi I. State of the art and practice in AI in education. European Journal of Education. 2022; 57; 4: 542-570. 10.1111/ejed.12533</bibtext> </blist> <blist> <bibtext> Hosseini S, Baziyad H, Norouzi R, Jabbedari Khiabani S, Gidófalvi G, Albadvi A, Alimohammadi A, Seyedabrishami S. Mapping the intellectual structure of GIS-T field (2008–2019): A dynamic co-word analysis. Scientometrics. 2021; 126: 2667-2688. 10.1007/s11192-020-03840-8</bibtext> </blist> <blist> <bibtext> Hoyle A, Goel P, Hian-Cheong A, Peskov D, Boyd-Graber J, Resnik P. Is automated topic model evaluation broken? The incoherence of coherence. Advances in Neural Information Processing Systems. 2021; 34: 2018-2033</bibtext> </blist> <blist> <bibtext> Huang AY, Lu OH, Yang SJ. Effects of artificial Intelligence-Enabled personalized recommendations on learners' learning engagement, motivation, and outcomes in a flipped classroom. Computers & Education. 2023; 194: 104684</bibtext> </blist> <blist> <bibtext> Hung J, Chen J. The benefits, risks and regulation of using ChatGPT in Chinese academia: A content analysis. Social Sciences. 2023; 12; 7: 380</bibtext> </blist> <blist> <bibtext> Hutto C, Gilbert E. VADER: A parsimonious rule-based model for sentiment analysis of social media text. Proceedings of the International AAAI Conference on Web and Social Media. 2014; 8; 1: 216-225</bibtext> </blist> <blist> <bibtext> Ifenthaler D, Yau JYK. Utilising learning analytics to support study success in higher education: A systematic review. Educational Technology Research and Development. 2020; 68; 4: 1961-1990. 10.1007/s11423-020-09788-z</bibtext> </blist> <blist> <bibtext> Jain R, Kumar A, Nayyar A, Dewan K, Garg R, Raman S, Ganguly S. Explaining sentiment analysis results on social media texts through visualization. Multimedia Tools and Applications. 2023; 82; 15: 22613-22629. 10.1007/s11042-023-14432-y</bibtext> </blist> <blist> <bibtext> Kajiwara Y, Matsuoka A, Shinbo F. Machine learning role playing game: Instructional design of AI education for age-appropriate in K-12 and beyond. Computers and Education. 2023; 5: 100162</bibtext> </blist> <blist> <bibtext> Khalil M, Er E. Will ChatGPT get you caught? Rethinking of plagiarism detection. International conference on human-computer interaction. 2023; Springer Nature Switzerland: 475-487</bibtext> </blist> <blist> <bibtext> Khowaja SA, Khuwaja P, Dev K, Wang W, Nkenyereye L. Chatgpt needs spade (sustainability, privacy, digital divide, and ethics) evaluation: A review. Cognitive Computation. 2024; 5: 1-23</bibtext> </blist> <blist> <bibtext> Kim J, Lee SS. Are two heads better than one? The effect of student-AI collaboration on students' learning task performance. TechTrends. 2023; 67; 2: 365-375. 10.1007/s11528-022-00788-9</bibtext> </blist> <blist> <bibtext> King CL, Vincent K, Warnars HL, Nordin N, Utomo WH. Intelligent tutoring system: Learning math for 6th-grade primary school students. Education Research International. 2021; 2021: 1-10</bibtext> </blist> <blist> <bibtext> Kordzadeh N, Ghasemaghaei M. Algorithmic bias: Review, synthesis, and future research directions. European Journal of Information Systems. 2022; 31; 3: 388-409</bibtext> </blist> <blist> <bibtext> Lin B, Cassee N, Serebrenik A, Bavota G, Novielli N, Lanza M. Opinion mining for software development: A systematic literature review. ACM Transactions on Software Engineering and Methodology (TOSEM). 2022; 31; 3: 1-41. 10.1145/3490388</bibtext> </blist> <blist> <bibtext> Long D, Magerko B. What is AI Literacy? Competencies and design considerations. Conference on Human Factors in Computing Systems. 2020. 10.1145/3313831.3376727</bibtext> </blist> <blist> <bibtext> Lund BD, Wang T. Chatting about ChatGPT: How may AI and GPT impact academia and libraries?. Library Hi Tech News. 2023; 40; 3: 26-29</bibtext> </blist> <blist> <bibtext> Ma S, Sun M, Ma X. Modeling adaptive expression of robot learning engagement and exploring its effects on human teachers. ACM Transactions on Computer-Human Interaction. 2023; 30; 5: 1-48</bibtext> </blist> <blist> <bibtext> Memarian B, Doleck T. Fairness, accountability, transparency, and ethics (FATE) in artificial intelligence (AI), and higher education: A systematic review. Computers and Education. 2023; 5: 100152</bibtext> </blist> <blist> <bibtext> Nanda G, Douglas KA, Waller DR, Merzdorf HE, Goldwasser D. Analyzing large collections of open-ended feedback from MOOC learners using LDA topic modeling and qualitative analysis. IEEE Transactions on Learning Technologies. 2021; 14; 2: 146-160</bibtext> </blist> <blist> <bibtext> Nguyen TH, Tran DN, Vo DL, Mai VH, Dao XQ. AI-powered university: Design and deployment of robot assistant for smart universities. Journal of Advances in Information Technology. 2022; 13; 1: 78</bibtext> </blist> <blist> <bibtext> Ouyang F, Wu M, Zheng L, Zhang L, Jiao P. Integration of artificial intelligence performance prediction and learning analytics to improve student learning in online engineering course. International Journal of Educational Technology in Higher Education. 2023; 20; 1: 4</bibtext> </blist> <blist> <bibtext> Ouyang F, Zheng L, Jiao P. Artificial intelligence in online higher education: A systematic review of empirical research from 2011 to 2020. Education and Information Technologies. 2022; 27; 6: 7893-7925. 10.1007/s10639-022-10925-9</bibtext> </blist> <blist> <bibtext> Pan X, Xue Y. Advancements of artificial intelligence techniques in the realm about library and information subject—A case survey of latent Dirichlet allocation method. IEEE Access. 2023; 11: 132627-132640</bibtext> </blist> <blist> <bibtext> Perrotta C, Selwyn N. Deep learning goes to school: Toward a relational understanding of AI in education. Learning, Media and Technology. 2020; 45; 3: 251-269</bibtext> </blist> <blist> <bibtext> Popenici S. The critique of AI as a foundation for judicious use in higher education. Journal of Applied Learning and Teaching. 2023; 6; 2: 1</bibtext> </blist> <blist> <bibtext> Pourhatami A, Kaviyani-Charati M, Kargar B, Baziyad H, Kargar M, Olmeda-Gómez C. Mapping the intellectual structure of the coronavirus field (2000–2020): A co-word analysis. Scientometrics. 2021; 126; 8: 6625-6657. 10.1007/s11192-021-04038-2</bibtext> </blist> <blist> <bibtext> Purgina M, Mozgovoy M, Blake J. WordBricks: Mobile technology and visual grammar formalism for gamification of natural language grammar acquisition. Journal of Educational Computing Research. 2020; 58; 1: 126-159</bibtext> </blist> <blist> <bibtext> Qin F, Li K, Yan J. Understanding user trust in artificial intelligence-based educational systems: Evidence from China. British Journal of Educational Technology. 2020; 51; 5: 1693-1710</bibtext> </blist> <blist> <bibtext> Rahman MM, Watanobe Y. ChatGPT for education and research: Opportunities, threats, and strategies. Applied Sciences. 2023; 13; 9: 5783</bibtext> </blist> <blist> <bibtext> Relmasira SC, Lai YC, Donaldson JP. Fostering AI literacy in elementary science, technology, engineering, art, and mathematics (STEAM) education in the age of generative AI. Sustainability. 2023; 15; 18: 13595. 10.3390/su151813595</bibtext> </blist> <blist> <bibtext> Reshi AA, Rustam F, Aljedaani W, Shafi S, Alhossan A, Alrabiah Z, Ahmad A, Alsuwailem H, Almangour TA, Alshammari MA, Lee E, Ashraf I. COVID-19 vaccination-related sentiments analysis: A case study using worldwide Twitter dataset. Healthcare. 2022; 10; 3: 411</bibtext> </blist> <blist> <bibtext> Shaik T, Tao X, Li Y, Dann C, McDonald J, Redmond P, Galligan L. A review of the trends and challenges in adopting natural language processing methods for education feedback analysis. IEEE Access. 2022; 10: 56720-56739</bibtext> </blist> <blist> <bibtext> Sun J, Ma H, Zeng Y, Han D, Jin Y. Promoting the AI teaching competency of K-12 computer science teachers: A TPACK-based professional development approach. Education and Information Technologies. 2023; 28; 2: 1509-1533</bibtext> </blist> <blist> <bibtext> Tahiru F. AI in education: A systematic literature review. Journal of Cases on Information Technology (JCIT). 2021; 23; 1: 1-20. 10.4018/JCIT.2021010101</bibtext> </blist> <blist> <bibtext> Touretzky D, Gardner-McCune C, Martin F, Seehorn D. Envisioning AI for K-12: What should every child know about AI?. Proceedings of the AAAI Conference on Artificial Intelligence. 2019; 33; 01: 9795-9799. 10.1609/aaai.v33i01.33019795</bibtext> </blist> <blist> <bibtext> Truong H, Qi D, Ryason A, Sullivan AM, Cudmore J, Alfred S, Jones SB, Parra JM, De S, Jones DB. Does your team know how to respond safely to an operating room fire? Outcomes of a virtual reality, AI-enhanced simulation training. Surgical Endoscopy. 2022; 36: 1-9</bibtext> </blist> <blist> <bibtext> Velander J, Taiye MA, Otero N, Milrad M. Artificial intelligence in K-12 education: Eliciting and reflecting on Swedish teachers' understanding of AI and its implications for teaching & learning. Education and Information Technologies. 2023. 10.1007/s10639-023-11990-4</bibtext> </blist> <blist> <bibtext> Wang T, Cheng EC. Towards a tripartite research agenda: A scoping review of artificial intelligence in education research. The international conference on artificial intelligence in education technology. 2021; Springer Nature Singapore: 3-24</bibtext> </blist> <blist> <bibtext> Wang Y. Exploration on the operation status and optimization strategy of networked teaching of physical education curriculum based on AI algorithm. International Journal of Information Technologies and Systems Approach (IJITSA). 2023; 16; 3: 1-15</bibtext> </blist> <blist> <bibtext> Whitehill, J, & Seltzer, M. (2017). A crowdsourcing approach to collecting tutorial videos—toward personalized learning-at-scale. In: Proceedings of the Fourth (2017) ACM Conference on Learning@ Scale (pp. 157–160).</bibtext> </blist> <blist> <bibtext> Williams, J. J, Kim, J, Rafferty, A, Maldonado, S, Gajos, K. Z, Lasecki, W. S, & Heffernan, N. (2016). Axis: Generating explanations at scale with learnersourcing and machine learning. In: Proceedings of the third (2016) ACM Conference on Learning@ Scale (pp. 379–388).</bibtext> </blist> <blist> <bibtext> Wolf RR, Wolf AB. Using AI to evaluate a competency-based online writing course in nursing. Online Learning. 2023; 27; 3: 41-69</bibtext> </blist> <blist> <bibtext> Wood EA, Ange BL, Miller DD. Are we ready to integrate artificial intelligence literacy into medical school curriculum: Students and faculty survey. Journal of Medical Education and Curricular Development. 2021; 8: 23821205211024078</bibtext> </blist> <blist> <bibtext> Wu D, Yang R, Shen C. Sentiment word co-occurrence and knowledge pair feature extraction based LDA short text clustering algorithm. Journal of Intelligent Information Systems. 2021; 56: 1-23. 10.1007/s10844-020-00597-7</bibtext> </blist> <blist> <bibtext> Xing W, Lee HS, Shibani A. Identifying patterns in students' scientific argumentation: Content analysis through text mining using latent Dirichlet allocation. Educational Technology Research and Development. 2020; 68; 5: 2185-2214. 10.1007/s11423-020-09761-w</bibtext> </blist> <blist> <bibtext> Zawacki-Richter O, Marín VI, Bond M, Gouverneur F. Systematic review of research on artificial intelligence applications in higher education—where are the educators?. International Journal of Educational Technology in Higher Education. 2019; 16; 1: 1-27</bibtext> </blist> <blist> <bibtext> Zhai X, Chu X, Chai CS, Jong MSY, Istenic A, Spector M, Liu J, Yuan J, Li Y. A review of Artificial Intelligence (AI) in education from 2010 to 2020. Complexity. 2021; 2021; 1: 8812542. 10.1155/2021/8812542</bibtext> </blist> <blist> <bibtext> Zou M, Huang L. The impact of ChatGPT on L2 writing and expected responses: Voice from doctoral students. Education and Information Technologies. 2023; 29: 1-19</bibtext> </blist> </ref> <ref id="AN0183751070-34"> <title> Footnotes </title> <blist> <bibtext> https://scholar.google.com/citations?view_op=top_venues&hl=en&vq=eng_educationaltechnology</bibtext> </blist> </ref> <aug> <p>By Ji Hyun Yu; Devraj Chauhan; Rubaiyat Asif Iqbal and Eugene Yeoh</p> <p>Reported by Author; Author; Author; Author</p> <p></p> <p>Ji Hyun Yu Ji Hyun Yu is an Assistant Professor of Learning Technologies at the College of Information, University of North Texas. She specializes in data-driven learning design, MOOCs, and AI in education, with a growing focus on developing emotionally intelligent AI tutors. Her research combines AI, learning analytics, and human-centered design to enhance personalized learning experiences, driven by a passion for improving educational outcomes through innovative technology.</p> <p>Devraj Chauhan Devraj Chauhan is a graduate student in data science at the University of North Texas, specializing in NLP and LDA techniques to personalize learning content. His expertise includes designing data engineering solutions and collaborating with analytics teams, gained from his experience as a Software Engineer at Nihilent.</p> <p>Rubaiyat Asif Iqbal Rubaiyat Asif Iqbal is a Ph.D. candidate in Learning Technologies at the University of North Texas, with a background in journalism, visual communication, and interaction design. His research focuses on Learning Experience Design (LXD) and the integration of Artificial Intelligence (AI) in Higher Education, aiming to enhance educational outcomes and promote design for social good.</p> <p>Eugene Yeoh Eugene Yeoh is a Ph.D. student in Learning Technologies at the University of North Texas. His research explores cognitive mechanisms behind learning difficulties in dementia, focusing on using technology to create virtual learning environments for neurorehabilitation and supporting life skills in this population.</p> </aug> <nolink nlid="nl1" bibid="bib38" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib14" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib29" firstref="ref4"></nolink> <nolink nlid="nl4" bibid="bib11" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib32" firstref="ref6"></nolink> <nolink nlid="nl6" bibid="bib35" firstref="ref7"></nolink> <nolink nlid="nl7" bibid="bib53" firstref="ref8"></nolink> <nolink nlid="nl8" bibid="bib69" firstref="ref9"></nolink> <nolink nlid="nl9" bibid="bib82" firstref="ref10"></nolink> <nolink nlid="nl10" bibid="bib59" firstref="ref17"></nolink> <nolink nlid="nl11" bibid="bib75" firstref="ref18"></nolink> <nolink nlid="nl12" bibid="bib76" firstref="ref19"></nolink> <nolink nlid="nl13" bibid="bib34" firstref="ref20"></nolink> <nolink nlid="nl14" bibid="bib44" firstref="ref21"></nolink> <nolink nlid="nl15" bibid="bib47" firstref="ref22"></nolink> <nolink nlid="nl16" bibid="bib67" firstref="ref23"></nolink> <nolink nlid="nl17" bibid="bib62" firstref="ref24"></nolink> <nolink nlid="nl18" bibid="bib48" firstref="ref26"></nolink> <nolink nlid="nl19" bibid="bib15" firstref="ref27"></nolink> <nolink nlid="nl20" bibid="bib28" firstref="ref28"></nolink> <nolink nlid="nl21" bibid="bib73" firstref="ref29"></nolink> <nolink nlid="nl22" bibid="bib18" firstref="ref30"></nolink> <nolink nlid="nl23" bibid="bib20" firstref="ref31"></nolink> <nolink nlid="nl24" bibid="bib24" firstref="ref32"></nolink> <nolink nlid="nl25" bibid="bib57" firstref="ref33"></nolink> <nolink nlid="nl26" bibid="bib81" firstref="ref34"></nolink> <nolink nlid="nl27" bibid="bib21" firstref="ref35"></nolink> <nolink nlid="nl28" bibid="bib56" firstref="ref36"></nolink> <nolink nlid="nl29" bibid="bib30" firstref="ref39"></nolink> <nolink nlid="nl30" bibid="bib16" firstref="ref40"></nolink> <nolink nlid="nl31" bibid="bib50" firstref="ref42"></nolink> <nolink nlid="nl32" bibid="bib31" firstref="ref43"></nolink> <nolink nlid="nl33" bibid="bib68" firstref="ref44"></nolink> <nolink nlid="nl34" bibid="bib72" firstref="ref45"></nolink> <nolink nlid="nl35" bibid="bib70" firstref="ref47"></nolink> <nolink nlid="nl36" bibid="bib41" firstref="ref50"></nolink> <nolink nlid="nl37" bibid="bib12" firstref="ref51"></nolink> <nolink nlid="nl38" bibid="bib36" firstref="ref59"></nolink> <nolink nlid="nl39" bibid="bib61" firstref="ref60"></nolink> <nolink nlid="nl40" bibid="bib79" firstref="ref61"></nolink> <nolink nlid="nl41" bibid="bib25" firstref="ref62"></nolink> <nolink nlid="nl42" bibid="bib49" firstref="ref63"></nolink> <nolink nlid="nl43" bibid="bib80" firstref="ref64"></nolink> <nolink nlid="nl44" bibid="bib13" firstref="ref65"></nolink> <nolink nlid="nl45" bibid="bib40" firstref="ref66"></nolink> <nolink nlid="nl46" bibid="bib42" firstref="ref67"></nolink> <nolink nlid="nl47" bibid="bib66" firstref="ref68"></nolink> <nolink nlid="nl48" bibid="bib37" firstref="ref69"></nolink> <nolink nlid="nl49" bibid="bib54" firstref="ref70"></nolink> <nolink nlid="nl50" bibid="bib22" firstref="ref71"></nolink> <nolink nlid="nl51" bibid="bib58" firstref="ref72"></nolink> <nolink nlid="nl52" bibid="bib26" firstref="ref73"></nolink> <nolink nlid="nl53" bibid="bib27" firstref="ref74"></nolink> <nolink nlid="nl54" bibid="bib46" firstref="ref75"></nolink> <nolink nlid="nl55" bibid="bib65" firstref="ref76"></nolink> <nolink nlid="nl56" bibid="bib77" firstref="ref79"></nolink> <nolink nlid="nl57" bibid="bib78" firstref="ref80"></nolink> <nolink nlid="nl58" bibid="bib55" firstref="ref81"></nolink> <nolink nlid="nl59" bibid="bib52" firstref="ref82"></nolink> <nolink nlid="nl60" bibid="bib43" firstref="ref83"></nolink> <nolink nlid="nl61" bibid="bib17" firstref="ref84"></nolink> <nolink nlid="nl62" bibid="bib71" firstref="ref87"></nolink> <nolink nlid="nl63" bibid="bib74" firstref="ref88"></nolink> <nolink nlid="nl64" bibid="bib23" firstref="ref92"></nolink> <nolink nlid="nl65" bibid="bib39" firstref="ref93"></nolink> <nolink nlid="nl66" bibid="bib60" firstref="ref94"></nolink> <nolink nlid="nl67" bibid="bib83" firstref="ref96"></nolink> <nolink nlid="nl68" bibid="bib63" firstref="ref101"></nolink> <nolink nlid="nl69" bibid="bib19" firstref="ref102"></nolink> <nolink nlid="nl70" bibid="bib10" firstref="ref105"></nolink> <nolink nlid="nl71" bibid="bib33" firstref="ref106"></nolink> <nolink nlid="nl72" bibid="bib51" firstref="ref107"></nolink> <nolink nlid="nl73" bibid="bib64" firstref="ref108"></nolink> <nolink nlid="nl74" bibid="bib45" firstref="ref110"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1462602
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Mapping Academic Perspectives on AI in Education: Trends, Challenges, and Sentiments in Educational Research (2018-2024)
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Ji+Hyun+Yu%22">Ji Hyun Yu</searchLink> (ORCID <externalLink term="http://orcid.org/0000-0001-9648-2582">0000-0001-9648-2582</externalLink>)<br /><searchLink fieldCode="AR" term="%22Devraj+Chauhan%22">Devraj Chauhan</searchLink><br /><searchLink fieldCode="AR" term="%22Rubaiyat+Asif+Iqbal%22">Rubaiyat Asif Iqbal</searchLink><br /><searchLink fieldCode="AR" term="%22Eugene+Yeoh%22">Eugene Yeoh</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Educational+Technology+Research+and+Development%22"><i>Educational Technology Research and Development</i></searchLink>. 2025 73(1):199-227.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 29
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2025
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Information Analyses
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Artificial+Intelligence%22">Artificial Intelligence</searchLink><br /><searchLink fieldCode="DE" term="%22Technology+Uses+in+Education%22">Technology Uses in Education</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Trends%22">Educational Trends</searchLink><br /><searchLink fieldCode="DE" term="%22Trend+Analysis%22">Trend Analysis</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Research%22">Educational Research</searchLink><br /><searchLink fieldCode="DE" term="%22Technology+Integration%22">Technology Integration</searchLink><br /><searchLink fieldCode="DE" term="%22Ethics%22">Ethics</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Policy%22">Educational Policy</searchLink><br /><searchLink fieldCode="DE" term="%22Policy+Formation%22">Policy Formation</searchLink><br /><searchLink fieldCode="DE" term="%22Evidence+Based+Practice%22">Evidence Based Practice</searchLink><br /><searchLink fieldCode="DE" term="%22Decision+Making%22">Decision Making</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1007/s11423-024-10425-2
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 1042-1629<br />1556-6501
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: How is the academic community conceptualizing and approaching the integration of AI in education, considering its potential, complexities, and challenges? This study addresses this fundamental question by employing a multifaceted approach that combines co-occurrence network analysis, latent Dirichlet allocation (LDA), and sentiment analysis on a corpus of abstracts from academic publications from 2018 to 2024. The findings reveal key themes in the scholarly discourse, including the centrality of ethical considerations, the impact of global events on AI adoption, and the practical applications of AI in educational management and policymaking. Moreover, the study identifies the main factors discussed in literature as influencing successful AI integration, the challenges and opportunities associated with AI in education, and the evolving academic perspectives on AI's role in educational settings. This comprehensive analysis of academic literature provides valuable insights into the current state of AI in education research, highlighting trends, challenges, and sentiments as they have evolved over time. By mapping the landscape of scholarly thought on this topic, this study aims to inform future research agendas, contribute to policy discussions, and provide a foundation for evidence-based decision-making in the development and implementation of AI technologies in educational contexts.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2025
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1462602
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1462602
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s11423-024-10425-2
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 29
        StartPage: 199
    Subjects:
      – SubjectFull: Artificial Intelligence
        Type: general
      – SubjectFull: Technology Uses in Education
        Type: general
      – SubjectFull: Educational Trends
        Type: general
      – SubjectFull: Trend Analysis
        Type: general
      – SubjectFull: Educational Research
        Type: general
      – SubjectFull: Technology Integration
        Type: general
      – SubjectFull: Ethics
        Type: general
      – SubjectFull: Educational Policy
        Type: general
      – SubjectFull: Policy Formation
        Type: general
      – SubjectFull: Evidence Based Practice
        Type: general
      – SubjectFull: Decision Making
        Type: general
    Titles:
      – TitleFull: Mapping Academic Perspectives on AI in Education: Trends, Challenges, and Sentiments in Educational Research (2018-2024)
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Ji Hyun Yu
      – PersonEntity:
          Name:
            NameFull: Devraj Chauhan
      – PersonEntity:
          Name:
            NameFull: Rubaiyat Asif Iqbal
      – PersonEntity:
          Name:
            NameFull: Eugene Yeoh
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 02
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 1042-1629
            – Type: issn-electronic
              Value: 1556-6501
          Numbering:
            – Type: volume
              Value: 73
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Educational Technology Research and Development
              Type: main
ResultId 1